Quantitative genetics
Part of a series on 
Genetics 

Key components 
History and topics 

Research 
Personalized medicine 
Personalized medicine 
Quantitative genetics is a branch of population genetics that deals with phenotypes that vary continuously (in characters such as height or mass)—as opposed to discretely identifiable phenotypes and geneproducts (such as eyecolour, or the presence of a particular biochemical).
Both branches use the frequencies of different alleles of a gene in breeding populations (gamodemes), and combine them with concepts from simple Mendelian inheritance to analyze inheritance patterns across generations and descendant lines. While population genetics can focus on particular genes and their subsequent metabolic products, quantitative genetics focuses more on the outward phenotypes, and makes summaries only of the underlying genetics.
Due to the continuous distribution of phenotypic values, quantitative genetics must employ many other statistical methods (such as the effect size, the mean and the variance) to link phenotypes (attributes) to genotypes. Some phenotypes may be analyzed either as discrete categories or as continuous phenotypes, depending on the definition of cutoff points, or on the metric used to quantify them.^{[1]}^{:27–69} Mendel himself had to discuss this matter in his famous paper,^{[2]} especially with respect to his peas attribute tall/dwarf, which actually was "length of stem".^{[3]}^{[4]} Analysis of quantitative trait loci, or QTL,^{[5]} is a more recent addition to quantitative genetics, linking it more directly to molecular genetics.
Contents

Basic principles 1
 Gene effects 1.1

Allele and genotype frequencies 1.2
 Random mating 1.2.1
 Mendel's research cross  a contrast 1.2.2
 Self fertilization  an alternative 1.2.3

Population mean 1.3
 The mean after random fertilization 1.3.1
 The mean after self fertilization 1.3.2
 The mean  generalized fertilization 1.3.3

Genetic drift 1.4
 The sample gamodemes  genetic drift 1.4.1
 The progeny lines  dispersion 1.4.2
 The equivalent postdispersion panmictic  inbreeding 1.4.3
 Extensive binomial sampling  is panmixia restored? 1.4.4

Continued genetic drift  increased dispersion and inbreeding 1.4.5
 Selfing within random fertilization 1.4.5.1
 Homozygosity and heterozygosity 1.4.6
 Allele shuffling  allele substitution 1.5

Extended principles 2
 Gene effects redefined 2.1
 Genotype substitution  expectations and deviations 2.2

Genotypic variance 2.3
 Genemodel approach  Mather Jinks Hayman 2.3.1
 Allelesubstitution approach  Fisher 2.3.2

Other fertilization patterns 2.4
 "Islands" random fertilization 2.4.1

Dispersion and the genotypic variance 2.5
 Derivation of s2G(1) 2.5.1
 Total dispersed genic variance  s2A(f) and af 2.5.2
 Total and partitioned dispersed dominance variances 2.5.3
 Environmental variance 2.6
 Heritability and repeatability 2.7

Relationship 3
 Pedigree analysis 3.1
 Crossmultiplication rules 3.2
 Fullsib crossing (FS) 3.3
 Halfsib crossing (HS) 3.4
 Self fertilization (SF) 3.5
 Cousins crossings 3.6
 Backcrossing (BC) 3.7

Relatedness between relatives 3.8
 Calculating relationship from known relationships 3.8.1
 Calculating relationship from pathway diagrams 3.8.2

Resemblances between relatives 4

Parentoffspring covariance 4.1
 Oneparent and offspring (PO) 4.1.1
 Midparent and offspring (MPO) 4.1.2
 Applications (parentoffspring) 4.1.3

Siblings covariances 4.2
 Halfsibs of the same commonparent (HS) 4.2.1
 Fullsibs (FS) 4.2.2
 Applications (siblings) 4.2.3

Parentoffspring covariance 4.1

Selection 5

Basic principles 5.1
 Alternative ?G 5.1.1

Background 5.2
 Standardized selection  the normal distribution 5.2.1
 Meiosis determination  reproductive path analysis 5.2.2

Dispersion and selection 5.3
 Coefficient of relationship as an intraclass correlation 5.3.1
 Phenotypic intraclass correlation 5.3.2

Relating intraclass correlations to dispersion and to Heritabilities 5.3.3
 Fundamentals 5.3.3.1
 Corollaries 5.3.3.2
 Amongstline selection 5.3.3.3
 Withinline selection (and overall individual selection) 5.3.3.4
 Selecting best individuals from best Lines  Combined selection 5.3.3.5
 Relative efficiencies of selection strategies 5.3.4
 Genetic drift and selection 5.4

Basic principles 5.1
 Correlated attributes 6
 See also 7
 Footnotes and references 8
 Further reading 9
 External links 10
Basic principles
Gene effects
In genotypic "value" (locus value) may be defined by the allele "effect" together with a dominance effect, and also by how genes interact with genes at other loci (epistasis). The founder of quantitative genetics  Sir Ronald Fisher  perceived much of this when he proposed the first mathematics of this branch of genetics.^{[6]}
Being a statistician, he defined the gene effects as deviations from a central value—enabling the use of statistical concepts such as mean and variance, which use this idea.^{[7]} The central value he chose for the gene was the midpoint between the two opposing homozygotes at the one locus. The deviation from there to the "greater" homozygous genotype can be named "+a" ; and therefore it is "a" from that same midpoint to the "lesser" homozygote genotype. This is the "allele" effect mentioned above. The heterozygote deviation from the same midpoint can be named "d", this being the "dominance" effect referred to above.^{[8]} The diagram depicts the idea. However, in reality we measure phenotypes, and the figure also shows how observed phenotypes relate to the gene effects. Formal definitions of these effects recognize this phenotypic focus.^{[9]}^{[10]} Epistasis has been approached statistically as interaction (i.e., inconsistencies),^{[11]} but epigenetics suggests a new approach may be needed.
If 0<d<a, the dominance is regarded as partial or incomplete—while d=a indicates full or classical dominance. Previously, d>a was known as "overdominance".^{[12]}
Mendel's pea attribute "length of stem" provides us with a good example.^{[3]} Mendel stated that the tall truebreeding parents ranged from 6–7 feet in stem length (183 – 213 cm), giving a median of 198 cm (= P1). The short parents ranged from 0.75  1.25 feet in stem length (23 – 46 cm), with a rounded median of 34 cm (= P2). Their hybrid ranged from 6–7.5 feet in length (183–229 cm), with a median of 206 cm (= F1). The mean of P1 and P2 is 116 cm, this being the phenotypic value of the homozygotes midpoint (mp). The allele affect (a) is [P1mp] = 82 cm = [P2mp]. The dominance effect (d) is [F1mp] = 90 cm.^{[13]} This historical example illustrates clearly how phenotype values and gene effects are linked.
Allele and genotype frequencies
To obtain means, variances and other statistics, both quantities and their occurrences are required. The gene effects (above) provide the framework for quantities: and the frequencies of the contrasting alleles in the fertilization gametepool provide the information on occurrences.
Commonly, the frequency of the allele causing "more" in the phenotype (including dominance) is given the symbol p, while the frequency of the contrasting allele is q. An initial assumption made when establishing the algebra was that the parental population was infinite and random mating, which was made simply to facilitate the derivation. The subsequent mathematical development also implied that the frequency distribution within the effective gametepool was uniform: there were no local perturbations where p and q varied. Looking at the diagrammatic analysis of sexual reproduction, this is the same as declaring that p_{P} = p_{g} = p; and similarly for q.^{[12]} This mating system, dependent upon these assumptions, became known as "panmixia".
Panmixia rarely actually occurs in nature,^{[14]}^{:152–180}^{[15]} as gamete distribution may be limited, for example by dispersal restrictions or by behaviour, or by chance sampling (those local perturbations mentioned above). It is wellknown that there is a huge wastage of gametes in Nature, which is why the diagram depicts a potential gametepool separately to the actual gametepool. Only the latter sets the definitive frequencies for the zygotes: this is the true "gamodeme" ("gamo" refers to the gametes, and "deme" derives from Greek for "population"). But, under Fisher's assumptions, the gamodeme can be effectively extended back to the potential gametepool, and even back to the parental basepopulation (the "source" population). The random sampling arising when small "actual" gametepools are sampled from a large "potential" gametepool is known as genetic drift, and is considered subsequently.
While panmixia may not be widely extant, the potential for it does occur, although it may be only ephemeral because of those local perturbations. It has been shown, for example, that the F2 derived from random fertilization of F1 individuals (an allogamous F2), following hybridization, is an origin of a new potentially panmictic population.^{[16]}^{[17]} It has also been shown that if panmictic random fertilization occurred continually, it would maintain the same allele and genotype frequencies across each successive panmictic sexual generation—this being the Hardy Weinberg equilibrium.^{[11]}^{:34–39}^{[18]}^{[19]}^{[20]}^{[21]} However, as soon as genetic drift was initiated by local random sampling of gametes, the equilibrium would cease.
Random mating
Male and female gametes within the actual fertilizing pool are considered usually to have the same frequencies for their corresponding alleles. (Exceptions have been considered.) This means that when p male gametes carrying the A allele randomly fertilize p female gametes carrying that same allele, the resulting zygote has genotype AA, and, under random fertilization, the combination occurs with a frequency of p x p (= p^{2}). Similarly, the zygote aa occurs with a frequency of q^{2}. Heterozygotes (Aa) can arise in two ways: when p male (A allele) randomly fertilize q female (a allele) gametes, and vice versa. The resulting frequency for the heterozygous zygotes is thus 2pq.^{[11]}^{:32} Notice that such a population is never more than half heterozygous, this maximum occurring when p=q= 0.5.
In summary then, under random fertilization, the zygote (genotype) frequencies are the quadratic expansion of the gametic (allelic) frequencies: ( p + q )^{2} = ( p^{2} + 2 p q + q^{2}) = 1. (The "=1" states that the frequencies are in fraction form, not percentages; and that there are no omissions within the framework proposed.) Also recall that "random fertilization" and "panmixia" are not synonyms, as discussed in the previous section.
Mendel's research cross  a contrast
Mendel's pea experiments were constructed by establishing truebreeding parents with "opposite" phenotypes for each attribute.^{[3]} This meant that each opposite parent was homozygous for its respective allele only. In our example, "tall vs dwarf", the tall parent would be genotype TT with p = 1 (and q = 0); while the dwarf parent would be genotype tt with q = 1 (and p = 0). After controlled crossing, their hybrid is Tt, with p = q = ½. However, the frequency of this heterozygote = 1, because this is the F1 of an artificial cross: it has not arisen through random fertilization.^{[22]} The F2 generation was produced by natural selfpollination of the F1 (with monitoring against insect contamination), resulting in p = q = ½ being maintained. Such an F2 is said to be "autogamous". However, the genotype frequencies (0.25 TT, 0.5 Tt, 0.25 tt) have arisen through a mating system very different from random fertilization, and therefore the use of the quadratic expansion has been avoided. The numerical values obtained were the same as those for random fertilization only because this is the special case of having originally crossed homozygous opposite parents.^{[23]} We can notice that, because of the dominance of T [frequency (0.25 + 0.5)] over tt [frequency 0.25], the 3:1 ratio is still obtained.
A cross such as Mendel's, where truebreeding (largely homozygous) opposite parents are crossed in a controlled way to produce an F1, is a special case of hybrid structure. The F1 is often regarded as "entirely heterozygous" for the gene under consideration. However, this is an oversimplification and does not apply generally—for example when individual parents are not homozygous, or when populations interhybridise to form hybrid swarms.^{[22]} The general properties of intraspecies hybrids (F1) and F2 (both "autogamous" and "allogamous") are considered in a later section.
Self fertilization  an alternative
Having noticed that the pea is naturally selfpollinated, we cannot continue to use it as an example for illustrating random fertilization properties. Selffertilization ("selfing") is a major alternative to random fertilization, especially within Plants. Most of the Earth's cereals are naturally selfpollinated (rice, wheat, barley, for example), as well as the pulses. Considering the millions of individuals of each of these on Earth at any time, it's obvious that selffertilization is at least as significant as random fertilization. Selffertilization is the most intensive form of inbreeding, which arises whenever there is restricted independence in the genetical origins of gametes. Such reduction in independence arises if parents are already related, and/or from genetic drift or other spatial restrictions on gamete dispersal. Path analysis demonstrates that these are tantamount to the same thing.^{[24]}^{[25]} Arising from this background, the inbreeding coefficient (often symbolized as F or f) quantifies the effect of inbreeding from whatever cause. There are several formal definitions of f, and some of these are considered in later sections. For the present, note that for a longterm selffertilized species f = 1. Natural selffertilized populations are not single " pure lines ", however, but mixtures of such lines. This becomes particularly obvious when considering more than one gene at a time. Therefore, allele frequencies (p and q) other than 1 or 0 are still relevant in these cases (refer back to the Mendel Cross section). The genotype frequencies take a different form, however.
In general, the genotype frequencies become [p^{2}(1f) + pf] for AA and [2pq(1f)] for Aa and [q^{2}(1f) + qf] for aa.^{[11]} ^{:65} Notice that the frequency of the heterozygote declines in proportion to f. When f = 1, these three frequencies become respectively p, 0 and q Conversely, when f = 0, they reduce to the randomfertilization quadratic expansion shown previously.
Population mean
The population mean shifts the central reference point from the homozygote midpoint (mp) to the mean of a sexually reproduced population. This is important not only to relocate the focus into the natural world, but also to use a measure of central tendency used by Statistics/Biometrics. In particular, the square of this mean is the Correction Factor, which is used to obtain the genotypic variances later.^{[7]}
For each genotype in turn, its allele effect is multiplied by its genotype frequency; and the products are accumulated across all genotypes in the model. Some algebraic simplification usually follows to reach a succinct result.
The mean after random fertilization
The contribution of AA is p^{2} (+)a, that of Aa is 2pq d. and that of aa is q^{2} ()a. Gathering together the two a terms and accumulating over all, the result is: a (p^{2}  q^{2}) + 2pq d. Simplification is achieved by noting that (p^{2}q^{2}) = (pq)(p+q), and by recalling that (p+q) = 1, thereby reducing the lefthand term to (pq). The succinct result is therefore G = a(pq) + 2pqd.^{[12]} ^{:110} This defines the population mean as an "offset" from the homozygote midpoint (recall a and d are defined as deviations from that midpoint). The Figure depicts G across all values of p for several values of d, including one case of slight overdominance. Notice that G is often negative, thereby emphasizing that it is itself a deviation (from mp). Finally, to obtain the actual Population Mean in "phenotypic space", the midpoint value is added to this offset: P = G + mp.
An example arises from data on ear length in maize.^{[26]}^{:103} Assuming for now that one gene only is represented, a = 5.45 cm, d = 0.12 cm [virtually "0", really], mp = 12.05 cm. Further assuming that p = 0.6 and q = 0.4 in this example population, then:
G = 5.45 (0.6  0.4) + (0.48)0.12 = 1.15 cm (rounded); and
P = 1.15 + 12.05 = 13.20 cm (rounded).
The mean after self fertilization
The contribution of AA is p (+)a, while that of aa is q ()a. (See above for the frequencies.) Gathering these two a terms together leads to an immediately very simple final result: G_{S} = a(pq). P is obtained as above.
Mendel's peas can provide us with the allele effects and midpoint (see previously); and a mixed selfpollinated population with p = 0.6 and q = 0.4 provides example frequencies. Thus:
G_{S} = 82 (0.6  .04) = 59.6 cm (rounded); and
P_{S} = 59.6 + 116 = 175.6 cm (rounded).
The mean  generalized fertilization
A general formula incorporates the inbreeding coefficient f, and can then accommodate any situation. The procedure is exactly the same as before, using the weighted genotype frequencies given earlier. After translation into our symbols, and further rearrangement:^{[11]} ^{:77–78}
G_{f} = a(qp) + [2pqd  f 2pqd] = a(qp) + (1f) 2pqd = G  f 2pqd
Supposing that the maize example (earlier) had been constrained on a holme (a narrow riparian meadow), and had partial inbreeding to the extent of f = 0.25, then, using the third version (above) of G_{f}:
G_{0.25} = 1.15  0.25 (0.48) 0.12 = 1.136 cm (rounded), with P = 13.194 cm (rounded).
There is hardly any effect from inbreeding in this example, which arises because there was virtually no dominance in this attribute (d > 0). Examination of all three versions of G_{f} reveals that this would lead to trivial change in the Population mean. Where dominance was notable, however, there would be considerable change.
Genetic drift
Genetic drift was introduced when discussing the likelihood of panmixia being widely extant as a natural fertilization pattern. [See section on Allele and Genotype frequencies.] Here the sampling of gametes from the potential gamodeme is discussed in more detail. The sampling involves random fertilization between pairs of random gametes, each of which may contain either an A or an a allele. The sampling is therefore binomial sampling.^{[11]}^{:382–395}^{[12]}^{:49–63}^{[27]}^{:35}^{[28]}^{:55} Each sampling "packet" involves 2N alleles, and produces N zygotes (a "progeny" or a "line") as a result. During the course of the reproductive period, this sampling is repeated over and over, so that the final result is a mixture of sample progenies. These events, and the overall endresult, are examined here with an illustrative example.
The "base" allele frequencies of the example are those of the potential gamodeme: the frequency of A is p_{g} = 0.75, while the frequency of a is q_{g} = 0.25. [White label "1" in the diagram.] Five example actual gamodemes are binomially sampled out of this base (s = the number of samples = 5), and each sample is designated with an "index" k: with k = 1 .... s sequentially. (These are the sampling "packets" referred to in the previous paragraph.) The number of gametes involved in fertilization varies from sample to sample, and is given as 2N_{k} [at white label "2" in the diagram]. The total (S) number of gametes sampled overall is 52 [white label "3" in the diagram]. Because each sample has its own size, weights are needed to obtain averages (and other statistics) when obtaining the overall results. These are ω_{k} = 2N_{k} / ( Σ_{k} 2N_{k} ), and are given at white label "4" in the diagram.
The sample gamodemes  genetic drift
Following completion of these five binomial sampling events, the resultant actual gamodemes each contained different allele frequencies—(p_{k} and q_{k}). [These are given at white label "5" in the diagram.] This outcome is actually the genetic drift itself. Notice that two samples (k = 1 and 5) happen to have the same frequencies as the base (potential) gamodeme. Another (k = 3) happens to have the p and q "reversed". Sample (k = 2) happens to be an "extreme" case, with p_{k} = 0.9 and q_{k} = 0.1 ; while the remaining sample (k = 4) is "middle of the range" in its allele frequencies. All of these results have arisen only by "chance", through binomial sampling. Having occurred, however, they set in place all the downstream properties of the progenies.
Because sampling involves chance, the probabilities ( ∫_{k} ) of obtaining each of these samples become of interest. These binomial probabilities depend on the starting frequencies (p_{g} and q_{g}) and the sample size (2N_{k}). They are tedious to obtain,^{[11]}^{:382–395}^{[28]}^{:55} but are of considerable interest. [See white label "6" in the diagram.] The two samples (k = 1, 5), with the allele frequencies the same as in the potential gamodeme, had higher "chances" of occurring than the other samples. Their binomial probabilities did differ, however, because of their different sample sizes (2N_{k}). The "reversal" sample (k = 3) had a very low Probability of occurring, confirming perhaps what might be expected. The "extreme" allele frequency gamodeme (k = 2) was not "rare", however; and the "middle of the range" sample (k=4) was rare. These same Probabilities apply also to the progeny of these fertilizations.
Here, some summarizing can begin. The overall allele frequencies in the progenies bulk are supplied by weighted averages of the appropriate frequencies of the individual samples. That is: p_{•} = [Σ_{k} (ω_{k} p_{k})] and q_{•} = [Σ_{k} (ω_{k} q_{k})]. (Notice that k is replaced by • for the overall result  a common practice.)^{[7]} The results for the example are p_{•} = 0.631 and q_{•} = 0.369 [black label "5" in the diagram]. These values are quite different to the starting ones (p_{g} and q_{g}) [white label "1"]. The sample allele frequencies also have variance as well as an average. This has been obtained using the sum of squares (SS) method ,^{[29]} [See to the right of black label "5" in the diagram]. [Further discussion on this variance occurs in the section below on Extensive genetic drift.]
The progeny lines  dispersion
The genotype frequencies of the five sample progenies are obtained from the usual quadratic expansion of their respective allele frequencies (random fertilization). The results are given at the diagram's white label "7" for the homozygotes, and at white label "8" for the heterozygotes. Rearrangement in this manner prepares the way for monitoring inbreeding levels. This can be done either by examining the level of total homozygosis [(p^{2}_{k} + q^{2}_{k}) = (1  2p_{k}q_{k})] , or by examining the level of heterozygosis (2p_{k}q_{k}), as they are complementary.^{[30]} Notice that samples k= 1, 3, 5 all had the same level of heterozygosis, despite one being the "mirror image" of the others with respect to allele frequencies. The "extreme" allelefrequency case (k= 2) had the most homozygosis (least heterozygosis) of any sample. The "middle of the range" case (k= 4) had the least homozygosity (most heterozygosity): they were each equal at 0.50, in fact.
The overall summary can continue by obtaining the weighted average of the respective genotype frequencies for the progeny bulk. Thus, for AA, it is p^{2}_{•} = [Σ_{k} (ω_{k} p^{2}_{k})] , for Aa , it is 2p_{•}q_{•} = [Σ_{k} (ω_{k} 2p_{k}q_{k})] and for aa, it is q^{2}_{•} = [Σ_{k} (ω_{k} q^{2}_{k})]. The example results are given at black label "7" for the homozygotes, and at black label "8" for the heterozygote. Note that the heterozygosity mean is 0.3588, which the next section uses to examine inbreeding resulting from this genetic drift.
The next focus of interest is the dispersion itself, which refers to the "spreading apart" of the progenies' population means. These are obtained as G_{k} = [a(p_{k}q_{k}) + 2p_{k}q_{k} d] [see section on the Population mean], for each sample progeny in turn, using the example gene effects given at white label "9" in the diagram. Then, each P_{k} = G_{k} + mp is obtained also. The latter are given at white label "10" in the diagram. Notice that the "best" line (k = 2) had the highest allele frequency for the "more" allele (A) (it also had the highest level of homozygosity). The worst progeny (k = 3) had the highest frequency for the "less" allele (a), which accounted for its poor performance. This "poor" line was less homozygous than the "best" line; and it shared the same level of homozygosity, in fact, as the two secondbest lines (k = 1, 5). The progeny line with both the "more" and the "less" alleles present in equal frequency (k = 4) had a mean below the overall average (see next paragraph), and had the lowest level of homozygosity. These results reveal the fact that the alleles most prevalent in the "genepool" (also called the "germplasm") determine performance, not the level of homozygosity per se. Binomial sampling alone effects this dispersion.
The overall summary can now be concluded by obtaining G_{•} = [Σ_{k} (ω_{k}G_{k})] and P_{•} = [Σ_{k} (ω_{k}P_{k})]. The example result for P_{•} is 36.94 (black label "10" in the diagram). This later quantifies inbreeding depression arising overall from the gamete sampling. However, recall that some "nondepressed" progeny means have been identified already (k = 1, 2, 5). This is an enigma of inbreeding  while there may be "depression" overall, there are usually superior lines among the gamodeme samplings.
The equivalent postdispersion panmictic  inbreeding
Included in the overall summary were the averaqe allele frequencies in the mixture of progeny lines (p_{•} and q_{•}). These can now be used to construct a hypothetical panmictic equivalent.^{[11]}^{:382–395}^{[12]}^{:49–63}^{[27]}^{:35} This can be regarded as a "reference" to assess the changes wrought by the gamete sampling. The example appends such a panmictic to the right of the Diagram. The frequency of AA is therefore (p_{•})^{2} = 0.3979. This is less than that found in the dispersed bulk (0.4513 at black label "7"). Similarly, for aa, (q_{•})^{2} = 0.1303  again less than the equivalent in the progenies bulk (0.1898). Clearly, genetic drift has increased the overall level of homozygosis by the amount (0.6411  0.5342) = 0.1069. In a complementary approach, the heterozygosity could be used instead. The panmictic equivalent for Aa is 2 p_{•} q_{•} = 0.4658, which is higher than that in the binomially sampled bulk (0.3588) [black label "8"]. The sampling has caused the heterozygosity to decrease by 0.1070, which differs trivially from the earlier estimate because of rounding errors.
The inbreeding coefficient (f) was introduced in the early section on Self Fertilization. Here, a formal definition of it is considered: f is the probability that two "same" alleles (that is A and A, or a and a), which fertilize together are of common ancestral origin  or (more formally) f is the probability that two homologous alleles are autozygous.^{[12]}^{[25]} Consider any random gamete in the potential gamodeme that has its syngamy partner restricted by binomial sampling. The probability that that second gamete is homologous autozygous to the first is 1/(2N), the reciprocal of the gamodeme size. For the five example progenies, these quantities are 0.1, 0.0833, 0.1, 0.0833 and 0.125 respectively, and their weighted average is 0.0961. This is the inbreeding coefficient of the example progenies bulk, provided it is unbiased with respect to the full binomial distribution. An example based upon s = 5 is likely to be biased, however, when compared to an appropriate entire binomial distribution based upon the sample number (s) approaching infinity (s → ∞). Another derived definition of f for the full Distribution is that f also equals the rise in homozygosity, which equals the fall in heterozygosity.^{[31]} For the example, these frequency changes are 0.1069 and 0.1070, respectively. This result is different to the above, indicating that bias with respect to the full underlying distribution is present in the example. For the example itself, these latter values are the better ones to use, namely f_{•} = 0.10695.
The population mean of the equivalent panmictic is found as [a (p_{•}q_{•}) + 2 p_{•}q_{•} d] + mp. Using the example gene effects (white label "9" in the diagram), this mean is 37.87. The equivalent mean in the dispersed bulk is 36.94 (black label "10"), which is depressed by the amount 0.93. This is the inbreeding depression from this Genetic Drift. However, as noted previously, three progenies were not depressed (k = 1, 2, 5), and had means even greater than that of the panmictic equivalent. These are the lines a plant breeder looks for in a line selection programme.^{[32]}
Extensive binomial sampling  is panmixia restored?
If the number of binomial samples is large (s → ∞ ), then p_{•} → p_{g} and q_{•} → q_{g}. It might be queried whether panmixia would effectively reappear under these circumstances. However, the sampling of allele frequencies has still occurred, with the result that σ^{2}_{p, q} ≠ 0. In fact, as s → ∞, the σ^{2}_{p, q} → [(p_{g}q_{g})/2N], which is the variance of the whole binomial distribution.^{[11]}^{:382–395}^{[12]}^{:49–63} Furthermore, the "Wahlund equations" show that the progenybulk homozygote frequencies can be obtained as the sums of their respective average values (p^{2}_{•} or q^{2}_{•}) and σ^{2}_{p, q}.^{[11]}^{:382–395} Likewise, the bulk heterozygote frequency is (2 p_{•} q_{•}) minus twice the σ^{2}_{p, q}. The variance arising from the binomial sampling is conspicuously present. Thus, even when s → ∞, the progenybulk genotype frequencies still reveal increased homozygosis, and decreased heterozygosis, there is still dispersion of progeny means, and still inbreeding and inbreeding depression. That is, panmixia is not reattained once lost—but a new potential panmixia can be initiated via an allogamous F2 following hybridization.^{[33]}
Continued genetic drift  increased dispersion and inbreeding
Previous discussion here on genetic drift examined just one cycle (generation) of the process. When the sampling continues over successive generations, conspicuous changes occur in s^{2}_{p}, q and f. Another "index" is needed to keep track of "time": t = 1 .... y where y = the number of "years" (generations) considered. The approach often is to add the current binomial increment (? = "de novo") to what has occurred previously.^{[11]} The entire binomial distribution is examined here. There is no further benefit from an abbreviated example.
First consider dispersion via the s^{2} _{p,q} . Earlier this variance was seen to be (p_{g} q_{g})/(2N) = p_{g} q_{g} [1 /(2N)] = p_{g} q_{g} f. [Notice particularly this last version, where f = [1 /(2N)] = ?f as the recurrences are followed.] With the extension over time, this is the result of the first cycle, so this can be designated as s^{2} _{1} (for brevity). At cycle 2, this variance is generated again, this time becoming the de novo variance, and accumulates to what was present already. The de novo is given the weight of "1", while the "carryover" is given the weight of 1[1/(2N)], which is the same as 1?f . Thus, s^{2}_{2} = 1 ?s^{2} + (1?f) s^{2}_{1} . After gathering terms, simplifying and recalling previous symbols, this becomes s^{2}_{2} = s^{2}_{1} {1 + [1?f]} . The extension to generalize to any time t , after considerable simplification, becomes: s^{2}_{t} = p_{g} q_{g} {1  [1  ?f]^{t} } .^{[11]}^{:328} Recall that it was the allele frequency variation that caused the "spreading apart" of the progenies' means (dispersion). Therefore, this s^{2}_{t} can be used to indicate the extent of dispersion over the generations.
The method for examining inbreeding is similar. The same weights as before are used respectively for de novo f ( ?f ) [recall this is 1/(2N) ] and carryover f. Therefore, f_{2} = 1 ?f + (1  ?f) f_{1} . In general, f_{t} = ?f + (1 ?f) f_{t1} = ?f (1 f_{t1} ) + f_{t1} after rearrangement.^{[11]} The graphs to the left show levels of inbreeding over twenty generations arising from genetic drift (binomial sampling of gamodemes) for various actual gamodeme sizes (2N).
Still further rearrangement of this general equation reveals some interesting relationships. After some simplification, (f_{t}  f_{t1}) = ?f (1f_{t1}). The lefthand side is the difference between the current and previous levels of inbreeding: the change in inbreeding (df_{t}). Notice, it is not ?f, unless the f_{t1} is 0. Another item of note is the (1f_{t1})  an index of noninbreeding  the panmictic index (p_{t1}). Further rearrangements reveal that ?f = df_{t} / p_{t1} = 1  [p_{t} / p_{t1}] , all useful relationships.
One further rearrangement gives f_{t} = 1  (1?f)^{t} (1f_{0}), which, assuming that f_{0} = 0, forms the section within the braces of the last equation above for the s^{2}_{t} . That is, s^{2}_{t} = p_{g} q_{g} f_{t} ! Rearranged, this provides also the revelation that f_{t} = s^{2}_{t} / (p_{g} q_{g}) . The two principal threads of binomial gamete sampling are thus cemented together, and are directly interchangeable.
Selfing within random fertilization
It is easy to overlook that random fertilization includes selffertilization. Sewall Wright showed that a proportion 1/N of random fertilizations is actually self fertilization, with the remainder (N1)/N being cross fertilization. Following path analysis and simplification, the divided random fertilization inbreeding was found to be: f_{t} = selfing_{t} + crossing_{t} = ?f (1+ f_{t1} ) + [(N1)/N] f_{t1} .^{[25]}^{[34]} Upon further rearrangement, the earlier results from the binomial sampling were confirmed, along with some new arrangements. Two of these were potentially very useful, namely: f_{t} = ?f [ 1 + f_{t1} (2N1)]; and further f_{t} = ?f(1f_{t1}) + f_{t1}).
The insight provided by this leads to some issues about the use of the inbreeding coefficient for binomial sampling random fertilization. Clearly, then, it is inappropriate for any species incapable of self fertilization, which includes plants with selfincompatibility mechanisms, dioecious plants, and bisexual animals (including, of course, mammals). The method Wright developed was modified to develop a random fertilization inbreeding equation that involved only cross fertilization with no self fertilization. The proportion 1/N formerly due to selfing now defined the carryover inbreeding arising from the cycle previous to the current cycle. The "inbreeding for cross fertilization only" final result was: f_{t} = f_{t1} + ?f (1 + f_{t2}  2 f_{t1}).^{[11]}^{:166} The graphs to the right depict the differences between standard random fertilization (based on binomial sampling, which includes selfing) RF, and binomial sampling adjusted for "cross fertilization alone" CF. As can be seen, the issue is nontrivial for small gamodeme sample sizes.
It now is necessary to note that not only is "panmixia" not a synonym for "random fertilization", but also that "random fertilization" is not a synonym for "cross fertilization".
Homozygosity and heterozygosity
In the section on the "Sample gamodemes  Genetic drift", a series of gamete samplings was followed. An important result of this sampling was that homozygosity rose at the expense of heterozygosity. This is one view of the genotype frequencies following the inbreeding attendant upon the sampling. Another view is linked to the definition of the inbreeding coefficient, and examines homozygotes according to whether they arose as allozygotes or autozygotes. Recall that autozygous alleles have the same allelic origin, the likelihood (frequency) of which is the inbreeding coefficient (f) by definition. The proportion arising allozygously is therefore (1f). For the Abearing gametes, which are present with a general frequency of p, the overall frequency of those that are autozygous is therefore (f p). Similarly, for abearing gametes, the autozygous frequency is (f q). [Remember that the issue of auto/allo zygosity can arise only for homologous alleles (that is A and A, or a and a), and not for nonhomologous alleles (A and a), which cannot possibly have the same allelic origin.] These two viewpoints regarding genotype frequencies must be connected to establish consistency.
Following firstly the auto/allo zygosity viewpoint, consider the allozygous component. This occurs with the frequency of (1f), and the alleles unite according to the random fertilization quadratic expansion. Thus: (1f)[(p_{0} + q_{0})^{2}] = (1f)[ p^{2}_{0} + q^{2}_{0} ] + (1f)[ 2 p_{0} q_{0} ] . Consider next the autozygous component. As these alleles are autozygous (same allelic origin), they are effectively selfings, and produce either AA or aa genotypes, but no heterozygotes. They therefore produce (f p_{0} + f q_{0}) genotypes, for AA and aa respectively. Adding the two components together results in: [(1f) p^{2}_{0} + f p_{0}] for the AA homozygote; [(1f) q^{2}_{0} + f q_{0}] for the aa homozygote; and (1f)[ 2 p_{0} q_{0} ] for the Aa heterozygote.^{[11]}^{:65}^{[12]} This is the same equation presented earlier in the section on "Self fertilization  an alternative". The reason for the decline in heterozygosity is made clear here. Heterozygotes can arise only from the allozygous component, and its frequency in the mix is just (1f)—the same as the frequency used for the heterozygotes.
Secondly, the sampling rise/fall viewpoint is followed. The decline in heterozygotes, f (2 p_{0} q_{0}) , is distributed equally towards each homozygote and added to their basic random fertilization expectations. Therefore, the frequencies are: (p^{2}_{0} + f p_{0} q_{0} ) for the AA homozygote; (q^{2}_{0} + f p_{0} q_{0} ) for the aa homozygote; and 2 p_{0} q_{0}  f (2 p_{0} q_{0}) for the heterozygote.
Thirdly, the consistency proof begins with considering the AA homozygote's final equation in the auto/allo zygosity paragraph above. Taking that equation—(1f) p^{2}_{0} + f p_{0} ]—expand the parentheses, and follow by regathering [within the resultant] the two new terms with the commonfactor f in them [resulting in: p^{2}_{0}  f ( p^{2}_{0}  p_{0} ) ]. Next, in the " p^{2}_{0} " in that previous parenthesis part, one of the p is substituted by (1q), becoming " (1q_{0} ) p_{0} " instead of " p^{2}_{0} ". Following that substitution, it is a straightforward matter of multiplyingout, simplifying and watching signs. The end result is " (p^{2}_{0} + f p_{0} q_{0} ) " , which is exactly the result for AA in the sampling rise/fall paragraph. The two viewpoints are therefore consistent for the AA homozygote. In a like manner, the consistency of the aa viewpoints can also be proven. For the heterozygote, the two viewpoints are in agreement from the beginning.
Allele shuffling  allele substitution
The genemodel examines the heredity pathway from the point of view of "inputs" (alleles/gametes) and "outputs" (genotypes/zygotes), with fertilization being the "process" converting one to the other. An alternative viewpoint concentrates on the "process" itself, and considers the zygote genotypes as arising from allele shuffling. In particular, it regards the results as if one allele had "substituted" for the other during the shuffle, together with a residual that deviates from this view. This formed an integral part of Fisher's method,^{[6]} in addition to his use of frequencies and effects to generate his genetical statistics.^{[12]}
A discursive derivation of the allele substitution alternative follows.^{[12]}^{:113} Suppose that the usual random fertilization of gametes in a "base" gamodeme  consisting of p gametes (A) and q gametes (a)  is replaced by fertilization with a "flood" of gametes all containing a single allele (A or a, but not both). The zygotic results can be interpreted in terms of the "flood" allele having "substituted for" the alternative allele in the underlying "base" gamodeme. The diagram assists in following this viewpoint: the upper part pictures an A substitution, while the lower part shows an a substitution. (The diagram's "RF allele" is the allele in the "base" gamodeme.)
Consider the upper part firstly. Because base A is present with a frequency of p, the substitute A fertilizes it with a frequency of p resulting in a zygote AA with an allele effect of a. Its contribution to the outcome, therefore, is the product p a. Similarly, when the substitute fertilizes base a (resulting in Aa with a frequency of q and heterozygote effect of d), the contribution is q d. The overall result of substitution by A is ,therefore, pa + qd. This is now oriented towards the population mean (see earlier section) by expressing it as a deviate from that mean : (pa + qd)  G. After some algebraic simplification, this becomes a _{A} = q [a + (qp)d]  the substitution effect of A.
A parallel reasoning can be applied to the lower part of the diagram, taking care with the differences in frequencies and gene effects. The result is the substitution effect of a, which is a _{a} = p [a + (qp)d].
The common factor inside the brackets is the average allele substitution effect, and is usually given as a = a + (qp)d.^{[12]}^{:113} It can also be derived in a more direct way, but the result is the same.
In subsequent sections, these substitution effects help define the genemodel genotypes as consisting of a partition predicted by these new effects (substitution expectations), and a deviation between these expectations and the previous genemodel effect. The expectations are also called the breeding value, and the deviations are also called dominance deviations. Ultimately, the variance arising from the substitution expectations becomes the additive genetic variance (s^{2}_{A})^{[12]} (also the genic variance^{[35]})—while that arising from the deviations becomes the dominance variance (s^{2}_{D}).
Extended principles
Gene effects redefined
The genemodel effects (a, d and a) are important soon in the derivation of the deviations from substitution expectations (d), which were first discussed in the previous Allele Substitution section. However, they need to be redefined themselves before they become useful in that exercise. They firstly need to be recentralized around the population mean (G), and secondly they need to be rearranged as functions of a, the average allele substitution effect.
The recentralized effect for AA, therefore, is a' = a  G which, after simplification, becomes a' = 2q(apd). The similar effect for Aa is d' = d  G = a(qp) + d(12pq), after simplification. Finally, the recentralized effect for aa is (a)' = 2p(a+qd).^{[12]}^{:116–119}
These recentralized effects eventually have the genotype substitution expectations (see next section) subtracted from them to subsequently defin the genotype substitution deviations. These expectations each are a function of the average allele substitution effect, and the present recentralized effects have to be rearranged still further to accommodate this last subtraction. Recalling that a = [a +(qp)d], rearrangement gives a = [a (qp)d]. After substituting this for a in a' and simplifying, the final version becomes a = 2q(aqd). Similarly, d' becomes d = a(qp) + 2pqd; and (a)' becomes (a) = 2p(a+pd).^{[12]}^{:118}
Genotype substitution  expectations and deviations
The zygote genotypes are the target of all this preparation. The homozygous genotype AA is a union of two substitution effects of A, one from each sex. Its substitution expectation is therefore a_{AA} = 2a_{A} = 2qa (see previous sections). Similarly, the substitution expectation of Aa is a_{Aa} = a_{A} + a_{a} = (qp)a ; and for aa, a_{aa} = 2a_{a} = 2pa. These substitution expectations of the genotypes also called breeding values.^{[12]}^{:114–116}
Substitution deviations are the differences between these expectations and the gene effects after their twostage redefinition in the previous section. Therefore, d_{AA} = a  a_{AA} = 2q^{2}d after simplification. Similarly, d_{Aa} = d  a_{Aa} = 2pqd after simplification. Finally, d_{aa} = (a)  a_{aa} = 2p^{2}d after simplification.^{[12]}^{:116–119}
The genotype substitution expectations ultimately give rise to the s^{2}_{A}, and the genotype substitution deviations give rise to the s^{2}_{D}.
Genotypic variance
There are two major approaches to defining and partitioning genotypic variance. One is based on the genemodel effects,^{[35]} while the other is based on the genotype substitution effects^{[12]} They are algebraically interconvertible with each other.^{[33]} In this section, the basic random fertilization derivation is considered, with the effects of inbreeding and dispersion set aside. This is dealt with later to arrive at a more general solution. Until this monogenic treatment is replaced by a multigenic one, and until epistasis is resolved in the light of the findings of epigenetics, the Genotypic variance has only the components considered here.
Genemodel approach  Mather Jinks Hayman
It is convenient to follow the Biometrical approach, which is based on correcting the unadjusted sum of squares (USS) by subtracting the correction factor (CF). Because all effects have been examined through frequencies, the USS can be obtained as the sum of the products of each genotype's frequency' and the square of its geneeffect. The CF in this case is the mean squared. The result is the SS, which, again because of the use of frequencies, is also immediately the variance.^{[7]}
The USS = p^{2}a^{2} + 2pqd^{2} + q^{2}(a)^{2} , and the CF = G^{2} . The SS = USS  CF = s^{2}_{G} .
After partial simplification,
s^{2}_{G} = 2pqa^{2} + (qp)4pqad + 2pqd^{2} + (2pq)^{2} d^{2} = s^{2}_{a} + (weighted_covariance)_{ad} + s^{2}_{d} + s^{2}_{D} = ½D + ½F´ + ½H_{1} + ¼H_{2} in Mather's terminology.^{[35]}^{:212}^{[36]}
Here, s^{2}_{a} represents the homozygote or allelic variance, and s^{2}_{d} represents the heterozygote or genemodel dominance variance. The randomfertilization dominance variance (s^{2}_{D}) is also present. These components are plotted across all values of p in the accompanying figure. Notice that the (weighted_covariance)_{ad}^{[37]} (hereafter abbreviated to cov_{ad}) is negative for 0.5
.
Further gathering of terms leads to ½D + ½F´ + ½H_{3} + ¼H_{2}, where ½H_{3} = (qp)^{2} ½H_{1} = (qp)^{2}2pqd^{2}. It is useful later in Diallel analysis, which is an experimental design for estimating these genetical statistics.^{[38]}
If, following the lastgiven rearrangements, the first three terms are amalgamated together, rearranged further and simplified, the result is the variance of the Fisherian substitution expectation. That is: s^{2}_{A} = s^{2}_{a} + cov_{ad} + s^{2}_{d}, a revealing insight indeed. Notice particularly that s^{2}_{A} is not s^{2}_{a}.^{[39]} Notice also that s^{2}_{D} = 2pq s^{2}_{d}. From the Figure, this can be visualized as accumulating s^{2}_{a}, s^{2}_{d} and cov to obtain s^{2}_{A}, while leaving the s^{2}_{D} still separated. It is clear also in the Figure that s^{2}_{D} < s^{2}_{d}, as expected from the equations.
The overall result is s^{2}_{G} = 2pq [a+(qp)d]^{2} + (2pq)^{2} d^{2} = s^{2}_{A} + s^{2}_{D} . However, its derivation via the substitution effects themselves is given in the next section.
Allelesubstitution approach  Fisher
Reference to the several earlier sections on allele substitution reveals that the two ultimate effects are genotype substitution expectations and genotype substitution deviations. Notice that these are each already defined as deviations from the random fertilization population mean (G). For each genotype in turn, the product of the frequency and the square of the relevant effect is obtained, and these are accumulated to obtain directly a SS and s^{2}. Details follow.
s^{2}_{A} = p^{2} a^{2}_{AA} + 2pq a^{2}_{Aa} + q^{2} a^{2}_{aa}, which simplifies to s^{2}_{A} = 2pqa^{2}.
s^{2}_{D} = p^{2} d^{2}_{AA} + 2pq d^{2}_{Aa} + q^{2} d^{2}_{aa}, which simplifies to s^{2}_{D} = (2pq)^{2} d^{2}.
Once again, s^{2}_{G} = s^{2}_{A} + s^{2}_{D} .
Note that this allelesubstitution approach defined the components separately, and then totaled them to obtain the final Genotypic variance. Conversely, the genemodel approach derived the whole situation (components and total) as one exercise. Bonuses arising from this were (a) the revelations about the real structure of s^{2}_{A}, and (b) the relative sizes of s^{2}_{d} and s^{2}_{D} (see previous subsection). It is also apparent that a "Mather" analysis is more informative, and that a "Fisher" analysis can always be constructed from it. The opposite conversion is not possible, however, because information about cov_{ad} would be missing.
Other fertilization patterns
In previous sections, dispersive random fertilization (genetic drift) has been considered comprehensively, and selffertilization and hybridizing have been examined to varying degrees. The diagram to the left depicts the first two of these, along with another "spatially based" pattern: islands. This is a pattern of random fertilization featuring dispersed gamodemes, with the addition of "overlaps" in which nondispersive random fertilization occurs. With the islands pattern, individual gamodeme sizes (2N) are observable, and overlaps (m) are minimal. This is one of Sewall Wright's array of possibilities.^{[34]} In addition to "spatially" based patterns of fertilization, there are others based on either "phenotypic" or "relationship" criteria. The phenotypic bases include assortative fertilization (between similar phenotypes) and disassortative fertilization (between opposite phenotypes). The relationship patterns include sib crossing, cousin crossing and backcrossing—and are considered in a separate section. Self fertilization may be considered both from a spatial or relationship point of view.
"Islands" random fertilization
The breeding population consists of s small dispersed random fertilization gamodemes of sample size 2N_{k} ( k = 1 ... s ) with " overlaps " of proportion m_{k} in which nondispersive random fertilization occurs. The dispersive proportion is thus (1  m_{k}) . The bulk population consists of weighted averages of sample sizes, allele and genotype frequencies and progeny means, as was done for genetic drift in an earlier section. However, each gamete sample size is reduced to allow for the overlaps, thus finding a 2N_{k} effective for (1  m_{k}) .
For brevity, the argument is followed further with the subscripts omitted. Recall that 1 / (2N) is ?f. Therefore, _{islands} ?f = (1  m)^{2} / [2N  (2N1)m^{2} ] . Notice that when m = 0 this reduces to the previous ?f. This is substituted into the inbreeding coefficient to obtain _{islands}f_{t} = _{islands}?f_{t} + ( 1  _{islands}?f_{t} ) _{islands}f_{(t1)} , where t is the index over generations, as before. The effective overlap proportion can be obtained also, as m_{t} = 1  [ ( 2N _{islands}?f_{t} ) / ((2N  1) _{islands}?f_{t} + 1 )] ^{(1/2)} . Here, the 2N refers to the unreduced sample size, not the islands adjustment.
The graphs to the right show the inbreeding for a gamodeme size of 2N = 50 for ordinary dispersed random fertilization (RF) (m=0), and for four overlap levels ( m = 0.0625, 0.125, 0.25, 0.5 ) of islands random fertilization. There has indeed been reduction in the inbreeding resulting from the nondispersed random fertilization in the overlaps: it is particularly notable as m appraoches 0.50. Sewall Wright suggested that this value should be the limit for the use of this approach.
Dispersion and the genotypic variance
In the section on genetic drift, and in other sections that discuss inbreeding, a major outcome from allele frequency sampling has been the dispersion of progeny means. This collection of means has its own average, and also has a variance: the amongstline variance. (This is a variance of the attribute itself, not of allele frequencies.) As dispersion develops further over succeeding generations, this amongstline variance would be expected to increase. Conversely, as homozygosity rises, the withinlines variance would be expected to decrease. The question arises therefore as to whether the total variance is changing—and, if so, in what direction. These issues are examined for both the genic (s^{2}_{A} ) and dominance ( s^{2}_{D} ) variances.
The crucial overview equation comes from Sewall Wright,^{[11]} ^{:99 & 130}^{[34]} and is the outline of the inbred genotypic variance based on a weighted average of its extremes, the weights being quadratic with respect to the inbreeding coefficient ( f ). This equation is:
s^{2}_{G(f)} = (1f) s^{2}_{G(0)} + f s^{2}_{G(1)} + f (1f) [ G_{0}  G_{1} ]^{2} ,
where f is the inbreeding coefficient, s^{2}_{G(0)} is the genotypic variance at f=0, s^{2}_{G(1)} is the genotypic variance at f=1, G_{0} is the population mean at f=0, and G_{1} is the population mean at f=1. The (1f) component concerns the reduction of variance within progeny lines. The f component addresses the increase in variance amongst progeny lines ; while the f (1f) component is seen (in the next line) to address a part of the dominance variance.^{[11]} ^{:99 & 130} These components can be expanded to reveal further insight.
Thus: s^{2}_{G(f)} = (1f) [ s^{2}_{A(0)} + s^{2}_{D(0)} ] + f (4pq a^{2} ) + f (1f) (2pq d)^{2} .
In the first component, s^{2}_{G(0)} has been expanded to show its two variance components as previously defined . The s^{2}_{G(1)} in the second component is 4pqa^{2} [which, recall, equals 2 (s^{2}_{a}) ] and is derived shortly. The third component's substitution is the result of the subtraction between the two "inbreeding extremes" of the population means (see section on the "Population Mean").^{[33]}
Summarising: the withinline components are (1f) s^{2}_{A(0)} and (1f) s^{2}_{D(0)} ; and the amongstline components are 2f s^{2}_{a(0)} and (f  f ^{2}) s^{2}_{D(0)}.^{[33]}
The total genic variance (additive genetic variance) is thus [ (1f) s^{2}_{A(0)} + 2f s^{2}_{a(0)} ] = (1+f) s^{2}_{A(f)} , where the s^{2}_{A(f)} is discussed shortly in a subsection. Similarly, the total dominance variance is thus [ (1f) s^{2}_{D(0)} + (f  f ^{2}) s^{2}_{D(0)} ] = (1f^{2}) s^{2}_{D(0)} .
Graphs to the left show these three genic variances, together with the three dominance variances, across all values of f, for p = 0.5 (at which the dominance variance is at a maximum). Graphs to the right show the Genotypic variance partitions (being the sums of the respective genic and dominance partitions) changing over ten generations with an example f = 0.10.
Answering, firstly, the questions posed at the beginning about the total variances [the S in the graphs] : the genetic variance rises linearly with the inbreeding coefficient, maximizing at twice its starting level. The dominance variance declines at the rate of (1  f^{2} ) (and therefore declines only slowly at low levels of inbreeding) until it finishes at zero. Now, notice the other trends. It is probably intuitive that the within line variances decline to zero with continued inbreeding, and this is seen to be the case [both at the same linear rate (1f) ]. The amongst line variances both increase with inbreeding up to f = 0.5, the genic variance at the rate of 2f, and the dominance variance at the rate of (f  f^{2}) . At 0.5 < f, however, the trends change. The amongst line genic variance continues its linear increase until it equals the total genic variance. But, the amongst line dominance variance now declines towards zero, because (f  f^{2}) also declines with 0.5 < f.^{[33]}
Derivation of s^{2}_{G(1)}
Recall that when f=1, heterozygosity is zero, withinline variance is zero, and all genotypic variance is thus amongstline variance and deplete of dominance variance. In other words, s^{2}_{G(1)} is the variance amongst fully inbred line means. Recall further (from "The mean after selffertilization" section) that such means (G_{1}'s, in fact) are G = a(pq). Substituting (1q) for the p, gives G_{1} = a (1 2q) = a  2aq.^{[12]}^{:265} Therefore, the s^{2}_{G(1)} is the s^{2}_{(a2aq)} actually. Now, in general, the variance of a difference (xy) is [ s^{2}_{x} + s^{2}_{y}  2 cov_{xy} ].^{[40]}^{:100}^{[41]} ^{:232} Therefore, s^{2}_{G(1)} = [ s^{2}_{a} + s^{2}_{2aq}  2 cov_{(a, 2aq)} ] . But a (an allele effect) and q (an allele frequency) are independent  so this covariance is zero. Furthermore, a is a constant from one line to the next, so s^{2}_{a} is also zero. Further, 2a is another constant (k), so the s^{2}_{2aq} is of the type s^{2}_{k X}. In general, the variance s^{2}_{k X} is equal to k^{2} s^{2}_{X} .^{[41]}^{:232} Putting all this together reveals that s^{2}_{(a2aq)} = (2a)^{2} s^{2}_{q} . Recall (from the section on "Continued genetic drift") that s^{2}_{q} = pq f . With f=1 here within this present derivation, this becomes pq 1 (that is pq), and it is substituted into the previous.
The final result is: s^{2}_{G(1)} = s^{2}_{(a2aq)} = 4a^{2} pq = 2(2pq a^{2}) = 2 s^{2}_{a} .
It follows immediately that f s^{2}_{G(1)} = f 2 s^{2}_{a} . [This last f comes from the initial Sewall Wright equation : it is not the f just set to "1" in the derivation concluded two lines above.]
Total dispersed genic variance  s^{2}_{A(f)} and a_{f}
Previous sections found that the within line genic variance is based upon the allele substitution "additive genetic variance" ( s^{2}_{A} )—but the amongst line genic variance is based upon the gene model "allele variance" ( s^{2}_{a} ). These two cannot simply be added to get total genic variance. [This is not a difficulty for the dominance variances because all components refer to the same base: s^{2}_{D}.] One approach to avoiding this problem was to revisit the derivation of the average allele substitution effect, and to construct a version, ( a _{f} ), that incorporates effects of the dispersion. Crow and Kimura achieved this^{[11]} ^{:130–131} using the recentered allele effects (a’, d’, (a)’ ) discussed previously ["Gene effects redefined"]. However, this was found subsequently to underestimate slightly the total Genic variance, and a new variancebased derivation led to a refined version.^{[33]}
The refined version is: a _{f} = { a^{2} + [(1f ) / (1 + f )] 2(q  p ) ad + [(1f ) / (1 + f )] (q  p )^{2} d^{2} } ^{(1/2)}
Consequently, s^{2}_{A(f)} = (1 + f ) 2pq a_{f} ^{2} does now agree with [ (1f) s^{2}_{A(0)} + 2f s^{2}_{a(0)} ] exactly.
Total and partitioned dispersed dominance variances
The total genic variance is of intrinsic interest in its own right. But it had had another important use as well: it was subtracted from Sewall Wright's inbred genotypic variance equation^{[34]} to provide an estimator for the (total) Dominance variance . An anomaly appeared, however, because the total dominance variance appeared to rise early in inbreeding despite the decline in heterozygosity.^{[12]} ^{:128} ^{:266} Consequently, the de novo derivation [referred to above] refined the equation for a_{f}.^{[33]} At the same time, a direct solution for the total dominance variance was obtained, thus avoiding the need for the "subtraction" method of previous times. Furthermore, by incorporating the expanded Sewall Wright equation, direct solutions for the dispersion partitions of the Dominance variance were obtained also.
Environmental variance
The environmental variance is phenotypic variability, which cannot be ascribed to genetics. This sounds simple, but the experimental design needed to separate the two needs very careful planning. Even the "external" environment can be divided into spatial and temporal components, as well as partitions such as "litter" or "family" and "culture" or "history". Where does epigenetic variance get placed? Is it embedded within epistasis: or is it "internal environment"? These components are very dependent upon the actual experimental model used to do the research. Such issues are very important when doing the research itself, but in this article on quantitative genetics this overview may suffice.
It is an appropriate place, however, for a summary:
Phenotypic variance = genotypic variances + environmental variances + genotypeenvironment interaction + experimental "error" variance
i.e., s²_{P} = s²_{G} + s²_{E} + s²_{GE} + s²
or s²_{P} = s²_{A} + s²_{D} + s²_{I} + s²_{E} + s²_{GE} + s²
after partitioning the genotypic variance (G) into the components of "additive" (A), "dominance" (D), and "epistasic" (I) variance mentioned above.^{[42]}
Heritability and repeatability
The heritability of a trait is the proportion of the total (phenotypic) variance (s²_{P}) that is explained by the total genotypic variance (s²_{G}). This is known as the "broad sense" heritability (H^{2}).^{[43]} If only additive genetic variance (s²_{A}) is used in the numerator, the heritability is called "narrow sense" (h^{2}).
The broad sense heritability indicates the proportion of the phenotypic variance due to the whole genotypical variance. In colloquial terms, it indicates the extent of "nature" while (1H^{2}) indicates the extent of "nurture". Narrow sense heritability indicates the proportion of the phenotypic variance attributable to the "additive" genetic variance, discussed above. It was pointed out there that this variance arises through substitution (i.e., phenotypic change) following fertilization. Fisher proposed that this narrowsense heritability might be appropriate in considering the results of natural selection, focusing as it does on changeability, and hence adaptation.^{[27]} It has been used also for predicting generally the results of artificial selection. In the latter case, however, the broad sense heritability may be more appropriate, as the whole attribute is being altered: not just adaptive capacity. Generally, advance from selection is more rapid with higher heritability. In animals, heritability of reproductive traits is typically low, while heritability of disease resistance and production are moderately low to moderate, and heritability of body conformation is high.
Repeatability (r^{2}) is the proportion of phenotypic variance attributable to differences in repeated measures of the same subject, arising from later records. It is used particularly for longlived species. This value can only be determined for traits that manifest multiple times in the organism's lifetime, such as adult body mass, metabolic rate or litter size. Individual birth mass, for example, would not have a repeatability value: but it would have a heritability value. Generally, but not always, repeatability indicates the upper level of the heritability.^{[44]}
r^{2} = (s²_{G} + s²_{PE})/s²_{P}
where s²_{PE} = phenotypeenvironment interaction = repeatability.
The above concept of repeatability is, however, problematic for traits that necessarily change greatly between measurements. For example, body mass increases greatly in many organisms between birth and adulthood. Nonetheless, within a given age range (or lifecycle stage), repeated measures could be done, and repeatability would be meaningful within that stage.
Relationship
From the heredity perspective, relations are individuals that inherited genes from one or more common ancestors. Therefore, their "relationship" can be quantified on the basis of the probability that they each have inherited a copy of an allele from the common ancestor. In earlier sections, the Inbreding coefficient has been defined as, "the probability that two same alleles ( A and A, or a and a ) have a common origin"—or, more formally, "The probability that two homologous alleles are autozygous." Previously, the emphasis was on an individual's likelihood of having two such alleles, and the coefficient was framed accordingly. It is obvious, however, that this probability of autozygosity for an individual must also be the probability that each of its two parents had this autozygous allele. In this refocused form, the probability is called the coancestry coefficient for the two individuals i and j ( f _{ij} ). In this form, it can be used to quantify relationship between two individuals, and may also be known as the coefficient of kinship or the consanguinity coefficient.^{[11]}^{:132–143} ^{[12]}^{:82–92}
Pedigree analysis
Pedigrees are diagrams of familial connections between individuals and their ancestors, and possibly between other members of the group that share genetical inheritance with them. They are relationship maps. A pedigree can be analyzed, therefore, to reveal coefficients of inbreeding and coancestry. Such pedigrees actually are informal depictions of path diagrams as used in path analysis, which was invented by Sewall Wright when he formulated his studies on inbreeding.^{[45]}^{:266–298} Using the diagram to the left, the probability that individuals "B" and "C" have received autozygous alleles from ancestor "A" is 1/2 (one out of the two diploid alleles). This is the "de novo" inbreeding (?f_{Ped}) at this step. However, the other allele may have had "carryover" autozygosity from previous generations, so the probability of this occurring is (de novo complement multiplied by the inbreeding of ancestor A ), that is (1  ?f_{Ped} ) f_{A} = (1/2) f_{A} . Therefore, the total probability of autozygosity in B and C, following the bifurcation of the pedigree, is the sum of these two components, namely (1/2) + (1/2)f_{A} = (1/2) (1+f _{A} ) . This can be viewed as the probability that two random gametes from ancestor A carry autozygous alleles, and in that context is called the coefficient of parentage ( f_{AA} ).^{[11]}^{:132–143}^{[12]}^{:82–92} It appears often in the following paragraphs.
Following the "B" path, the probability that any autozygous allele is "passed on" to each successive parent is again (1/2) at each step (including the last one to the "target" X ). The overall probability of transfer down the "B path" is therefore (1/2)^{3} . The power that (1/2) is raised to can be viewed as "the number of intermediates in the path between A and X ", n_{B} = 3 . Similarly, for the "C path", n_{C} = 2 , and the "transfer probability" is (1/2)^{2} . The combined probability of autozygous transfer from A to X is therefore [ f_{AA} (1/2)^{(nB)} (1/2)^{(nC)} ] . Recalling that f_{AA} = (1/2) (1+f _{A} ) , f_{X} = f_{PQ} = (1/2)^{(nB + nC + 1)} (1 + f_{A} ) . In this example, assuming that f_{A} = 0, f_{X} = 0.0156 (rounded) = f_{PQ} , one measure of the "relatedness" between P and Q.
Crossmultiplication rules
In the following sections on sibcrossing and similar topics, a number of "averaging rules" are useful. These derive from path analysis.^{[45]} The rules show that any coancestry coefficient can be obtained as the average of crossover coancestries between appropriate grandparental and parental combinations. Thus, referring to the diagram to the right, Crossmultiplier 1 is that f_{PQ} = average of ( f_{AC} , f_{AD} , f_{BC} , f_{BD} ) = (1/4) [f_{AC} + f_{AD} + f_{BC} + f_{BD} ] = f_{Y} also. In a similar fashion, crossmultiplier 2 states that f_{PC} = (1/2) [ f_{AC} + f_{BC} ]—while crossmultiplier 3 states that f_{PD} = (1/2) [ f_{AD} + f_{BD} ] . Returning to the first multiplier, it can now be seen also to be f_{PQ} = (1/2) [ f_{PC} + f_{PD} ], which, after substituting multipliers 2 and 3, resumes its original form.
In much of the following, the grandparental generation is referred to as (t2) , the parent generation as (t1) , and the "target" generation as t.
Fullsib crossing (FS)
The diagram to the right shows that full sib crossing is a direct application of crossMultiplier 1, with the slight modification that parents A and B repeat (in lieu of C and D) to indicate that individuals P1 and P2 have both of their parents in common—that is they are full siblings. Individual Y is the result of the crossing of two full siblings. Therefore, f_{Y} = f_{P1,P2} = (1/4) [ f_{AA} + 2 f_{AB} + f_{BB} ] . Recall that f_{AA} and f_{BB} were defined earlier (in Pedigree analysis) as coefficients of parentage, equal to (1/2)[1+f_{A} ] and (1/2)[1+f_{B} ] respectively, in the present context. Recognize that, in this guise, the grandparents A and B represent generation (t2) . Thus, assuming that in any one generation all levels of inbreeding are the same, these two coefficients of parentage each represent (1/2) [1 + f_{(t2)} ] .
Now, examine f_{AB} . Recall that this also is f_{P1} or f_{P2} , and so represents their generation  f_{(t1)} . Putting it all together, f_{t} = (1/4) [ 2 f_{AA} + 2 f_{AB} ] = (1/4) [ 1 + f_{(t2)} + 2 f_{(t1)} ] . That is the inbreeding coefficient for FullSib crossing .^{[11]}^{:132–143}^{[12]}^{:82–92} The graph to the left shows the rate of this inbreeding over twenty repetitive generations. The "repetition" means that the progeny after cycle t become the crossing parents that generate cycle (t+1 ), and so on successively. The graphs also show the inbreeding for random fertilization 2N=20 for comparison. Recall that this inbreeding coefficient for progeny Y is also the coancestry coefficient for its parents, and so is a measure of the relatedness of the two Fill siblings.
Halfsib crossing (HS)
Derivation of the half sib crossing takes a slightly different path to that for Full sibs. In the diagram to the right, the two halfsibs at generation (t1) have only one parent in common  parent "A" at generation (t2). The crossmultiplier 1 is used again, giving f_{Y} = f_{(P1,P2)} = (1/4) [ f_{AA} + f_{AC} + f_{BA} + f_{BC} ] . There is just one coefficient of parentage this time, but three coancestry coefficients at the (t2) level (one of them  f_{BC}  being a "dummy" and not representing an actual individual in the (t1) generation). As before, the coefficient of parentage is (1/2)[1+f_{A} ] , and the three coancestries each represent f_{(t1)} . Recalling that f_{A} represents f_{(t2)} , the final gathering and simplifying of terms gives f_{Y} = f_{t} = (1/8) [ 1 + f_{(t2)} + 6 f_{(t1)} ] .^{[11]}^{:132–143}^{[12]}^{:82–92} The graphs at left include this halfsib (HS) inbreeding over twenty successive generations.
As before, this also quantifies the relatedness of the two halfsibs at generation (t1) in its alternative form of f_{(P1, P2)} .
Self fertilization (SF)
A pedigree diagram for selfing is on the right. It is so straightforward it doesn't require any crossmultiplication rules. It employs just the basic juxtaposition of the inbreeding coefficient and its alternative the coancestry coefficient; followed by recognizing that, in this case, the latter is also a coefficient of parentage. Thus, f_{Y} = f_{(P1, P1)} = f_{t} = (1/2) [ 1 + f_{(t1)} ] .^{[11]}^{:132–143}^{[12]}^{:82–92} This is the fastest rate of inbreeding of all types, as can be seen in the graphs above. The selfing curve is, in fact, a graph of the coefficient of parentage.
Cousins crossings
These are derived with methods similar to those for siblings.^{[11]}^{:132–143}^{[12]}^{:82–92} As before, the coancestry viewpoint of the inbreeding coefficient provides a measure of "relatedness" between the parents P1 and P2 in these cousin expressions.
The pedigree for First Cousins (FC) is given to the right. The prime equation is f_{Y} = f_{t} = f_{P1,P2} = (1/4) [ f_{1D} + f_{12} + f_{CD} + f_{C2} ]. After substitution with corresponding inbreeding coefficients, gathering of terms and simplifying, this becomes f_{t} = (1/4) [ 3 f_{(t1)} + (1/4) [2 f_{(t2)} + f_{(t3)} + 1 ]] , which is a version for iteration  useful for observing the general pattern, and for computer programming. A "final" version is f_{t} = (1/16) [ 12 f_{(t1)} + 2 f_{(t2)} + f_{(t3)} + 1 ] .
The Second Cousins (SC) pedigree is on the left. Parents in the pedigree not related to the common Ancestor are indicated by numerals instead of letters. Here, the prime equation is f_{Y} = f_{t} = f_{P1,P2} = (1/4) [ f_{3F} + f_{34} + f_{EF} + f_{E4} ]. After working through the appropriate algebra, this becomes f_{t} = (1/4) [ 3 f_{(t1)} + (1/4) [3 f_{(t2)} + (1/4) [2 f_{(t3)} + f_{(t4)} + 1 ]]] , which is the iteration version. A "final" version is f_{t} = (1/64) [ 48 f_{(t1)} + 12 f_{(t2)} + 2 f_{(t3)} + f_{(t4)} + 1 ] .
To visualize the pattern in full cousin equations, start the series with the full sib equation rewritten in iteration form: f_{t} = (1/4)[2 f_{(t1)} + f_{(t2)} + 1 ]. Notice that this is the "essential plan" of the last term in each of the cousin iterative forms: with the small difference that the generation indices increment by "1" at each cousin "level". Now, define the cousin level as k = 1 (for First cousins), = 2 (for Second cousins), = 3 (for Third cousins), etc., etc.; and = 0 (for Full Sibs, which are "zero level cousins"). The last term can be written now as: (1/4) [ 2 f_{(t(1+k))} + f_{(t(2+k))} + 1] . Stacked in front of this last term are one or more iteration increments in the form (1/4) [ 3 f_{(tj)} + ... , where j is the iteration index and takes values from 1 ... k over the successive iterations as needed. Putting all this together provides a general formula for all levels of full cousin possible, including Full Sibs. For kth level full cousins, f{k}_{t} = ?ter_{j = 1}^{k} { (1/4) [ 3 f_{(tj)} + }_{j} + (1/4) [ 2 f_{(t(1+k))} + f_{(t(2+k))} + 1] . At the commencement of iteration, all f_{(tx)} are set at "0", and each has its value substituted as it is calculated through the generations. The graphs to the right show the successive inbreeding for several levels of Full Cousins.
For first halfcousins (FHC), the pedigree is to the left. Notice there is just one common ancestor (individual A). Also, as for second cousins, parents not related to the common ancestor are indicated by numerals. Here, the prime equation is f_{Y} = f_{t} = f_{P1,P2} = (1/4) [ f_{3D} + f_{34} + f_{CD} + f_{C4} ]. After working through the appropriate algebra, this becomes f_{t} = (1/4) [ 3 f_{(t1)} + (1/8) [6 f_{(t2)} + f_{(t3)} + 1 ]] , which is the iteration version. A "final" version is f_{t} = (1/32) [ 24 f_{(t1)} + 6 f_{(t2)} + f_{(t3)} + 1 ] . The iteration algorithm is similar to that for full cousins, except that the last term is (1/8) [ 6 f_{(t(1+k))} + f_{(t(2+k))} + 1 ] . Notice that this last term is basically similar to the half sib equation, in parallel to the pattern for full cousins and full sibs. In other words, half sibs are "zero level" half cousins.
There is a tendency to regard cousin crossing with a humanoriented point of view, possibly because of a wide interest in Genealogy. The use of pedigrees to derive the inbreeding perhaps reinforces this "Family History" view. However, such kinds of intercrossing occur also in natural populations—especially those that are sedentary, or have a "breeding area" that they revisit from season to season. The progenygroup of a harem with a dominant male, for example, may contain elements of sibcrossing, cousin crossing, and backcrossing, as well as genetic drift, especially of the "island" type. In addition to that, the occasional "outcross" adds an element of hybridization to the mix. It is not panmixia.
Backcrossing (BC)
Following the hybridizing between A and R, the F1 (individual B) is crossed back (BC1) to an original parent (R) to produce the BC1 generation (individual C). [It is usual to use the same label for the act of making the backcross and for the generation produced by it. The act of backcrossing is here in italics. ] Parent R is the recurrent parent. Two successive backcrosses are depicted, with individual D being the BC2 generation. These generations have been given t indices also, as indicated. As before, f_{D} = f_{t} = f_{CR} = (1/2) [ f_{RB} + f_{RR} ] , using crossmultiplier 2 previously given. The f_{RB} just defined is the one that involves generation (t1) with (t2). However, there is another such f_{RB} contained wholly within generation (t2) as well, and it is this one that is used now: as the coancestry of the parents of individual C in generation (t1). As such, it is also the inbreeding coefficient of C, and hence is f_{(t1)}. The remaining f_{RR} is the coefficient of parentage of the recurrent parent, and so is (1/2) [1 + f_{R} ] . Putting all this together : f_{t} = (1/2) [ (1/2) [ 1 + f_{R} ] + f_{(t1)} ] = (1/4) [ 1 + f_{R} + 2 f_{(t1)} ] . The graphs at right illustrate Backcross inbreeding over twenty backcrosses for three different levels of (fixed) inbreeding in the Recurrent parent.
This routine is commonly used in Animal and Plant Breeding programmes. Often after making the hybrid (especially if individuals are shortlived), the recurrent parent needs separate "line breeding" for its maintenance as a future recurrent parent in the backcrossing. This maintenance may be through selfing, or through fullsib or halfsib crossing, or through restricted randomlyfertilized populations, depending on the species' reproductive possibilities. Of course, this incremental rise in f_{R} carriesover into the f_{t} of the backcrossing. The result is a more gradual curve rising to the asymptotes than shown in the present graphs, because the f_{R} is not at a fixed level from the outset.
Relatedness between relatives
Central in estimating the variances for the various components is the principle of relatedness. A child has a father and a mother. Consequently, the child and father share 50% of their alleles, as do the child and the mother. However, the mother and father normally do not share alleles as a result of shared ancestors. Similarly, two full siblings share also on average 50% of the alleles with each other, while half siblings share only 25% of their alleles. This variation in relatedness can be used to estimate which proportion of the total phenotypic variance (s²_{P}) the abovementioned components explain.
The principle of relationship (R) is central to understanding the resemblances within families and can be useful when calculating inbreeding. Relationship has two definitions that can be applied: The probable portion of genes that are the same for two individuals due to common ancestry exceeding that of the base population Additive/numerator relationship: the relationship coefficient (Rxy¬) = twice the probability of two genes at loci in different individuals being identical by descent. Rxy values can range from 0 to 1. Relationship can be calculated in several ways; from the known relationships of the individual, from bracket pedigrees, and from pedigree path diagrams.
Calculating relationship from known relationships
Relationship  Relationship Coefficient 

Individual and itself  1.00 
Individual and a monozygotic twin  1.00 
Individual and parent  0.50 
Full siblings  0.50 
Half siblings  0.25 
Individual and grandparent  0.25 
Son of sire and daughter of sire  0.125 
Grandson and granddaughter of sire  0.0625 
 Note: if the common ancestor is inbred, multiply the relationship by (1+inbreeding coefficient)
Calculating relationship from pathway diagrams
RXY = S(.5)n(1+FCA)
n = number of segregations between X and Y through their common ancestor FCA = the inbreeding coefficient of the common ancestor
Example: calculating RAE and RBE Note: valid pathways only go through ancestors (only go against the direction of the arrow). For example, to calculate the relationship of A and B, the pathway ADB would be acceptable, whereas the pathway AXB would be not. The reason behind this is that having progeny together does not make two individuals related.
RAB: there are two possible pathways from A to E. ADFE = (1/2)3 = .125 ADE = (1/2)2 = .25 Total: .375
RBE: there are four possible pathways from B to E. BDE = (1/2)2 = .25 BDFE = (1/2)3 = .125 BCDE = (1/2)3 = .125 BCDFE = (1/2)4 = .0625 Total: .5625
The square root of h^2 equals the correlation between additive genotype and expressed phenotype, as shown through the general procedures of Path Analysis.^{[45]}^{:214–298}
Resemblances between relatives
These, in like manner to the Genotypic variances, can be derived through either the genemodel ("Mather") approach or the allelesubstitution ("Fisher") approach. Here, each method is demonstrated for alternate cases.
Parentoffspring covariance
These can be viewed either as the covariance between any offspring and any one of its parents (PO), or as the covariance between any offspring and the "midparent" value of both its parents (MPO).
Oneparent and offspring (PO)
This can be derived as the sum of crossproducts between parent geneeffects and onehalf of the progeny expectations using the allelesubstitution approach. The onehalf of the progeny expectation accounts for the fact that only one of the two parents is being considered. The appropriate parental geneeffects are therefore the secondstage redefined gene effects used to define the Genotypic variances earlier, that is: a = 2q(a  qd) and d = (qp)a + 2pqd and also (a) = 2p(a + pd) [see section "Gene effects redefined"]. Similarly, the appropriate progeny effects, for allelesubstitution expectations are onehalf of the earlier breeding values, the latter being: a_{AA} = 2qa, and a_{Aa} = (qp)a and also a_{aa} = 2pa [see section on "Genotype substitution  Expectations and Deviations"].
Because all of these effects are defined already as deviates from the genotypic mean, the crossproduct sum using {genotypefrequency * parental geneeffect * halfbreedingvalue} immediately provides the allelesubstitutionexpectation covariance between any one parent and its offspring. After careful gathering of terms and simplification, this becomes cov(PO)_{A} = pqa^{2} = ½ s^{2}_{A} .^{[11]} ^{:132–141}^{[12]} ^{:134–147}
Unfortunately, the allelesubstitutiondeviations are usually overlooked, but they have not "ceased to exist" nonetheless! Recall that these deviations are: d_{AA} = 2q^{2} d, and d_{Aa} = 2pq d and also d_{aa} = 2p^{2} d [see section on "Genotype substitution  Expectations and Deviations"]. Consequently, the crossproduct sum using {genotypefrequency * parental geneeffect * halfsubstitutiondeviations} also immediately provides the allelesubstitutiondeviations covariance between any one parent and its offspring. Once more, after careful gathering of terms and simplification, this becomes cov(PO)_{D} = 2p^{2}q^{2}d^{2} = ½ s^{2}_{D} .
It follows therefore that: cov(PO) = cov(PO)_{A} + cov(PO)_{D} = ½ s^{2}_{A} + ½ s^{2}_{D} , when dominance is not overlooked !
Midparent and offspring (MPO)
Because there are many combinations of parental genotypes, there are many different midparents and offspring means to consider, together with the varying frequencies of obtaining each parental pairing. The genemodel approach is the most expedient in this case. Therefore, an unadjusted sum of crossproducts (USCP)—using all products { parentpairfrequency * midparentgeneeffect * offspringgenotypemean }—is adjusted by subtracting the {overall genotypic mean}^{2} as correction factor (CF). After multiplying out all the various combinations, carefully gathering terms, simplifying, factoring and cancellingout where applicable, this becomes:
cov(MPO) = pq [a + (qp)d ]^{2} = pq a^{2} = ½ s^{2}_{A} , with no dominance having been overlooked in this case, as it had been usedup in defining the a.^{[11]} ^{:132–141}^{[12]} ^{:134–147}
Applications (parentoffspring)
The most obvious application is an experiment that contains all parents and their offspring, with or without reciprocal crosses, preferably replicated without bias, enabling estimation of all appropriate means, variances and covariances, together with their standard errors. These estimated statistics can then be used to estimate the genetical variances. Twice the difference between the estimates of the two forms of (corrected) parentoffspring covariance provides an estimate of s^{2}_{D}; and twice the cov(MPO) estimates s^{2}_{A}. With appropriate experimental design and analysis,^{[7]}^{[40]}^{[41]} standard errors can be obtained for these genetical statistics as well. This is the basic core of an experiment known as Diallel analysis, the Mather, Jinks and Hayman version of which is discussed in another section.
A second application involves using regression analysis, which estimates from statistics the ordinate (Yestimate), derivative (regression coefficient) and constant (Yintercept) of calculus.^{[7]}^{[40]}^{[46]}^{[47]} The regression coefficient estimates the rate of change of the function predicting Y from X, based on minimizing the residuals between the fitted curve and the observed data (MINRES). No alternative method of estimating such a function satisfies this basic requirement of MINRES. In general, the regression coefficient is estimated as the ratio of the covariance(XY) to the variance of the determinator (X). In practice, the sample size is usually the same for both X and Y, so this can be written as SCP(XY) / SS(X), where all terms have been defined previously.^{[7]}^{[46]}^{[47]} In the present context, the parents are viewed as the "determinative variable" (X), and the offspring as the "determined variable" (Y), and the regression coefficient as the "functional relationship" (ß_{PO}) between the two. Taking cov(MPO) = ½ s^{2}_{A} as cov(XY), and s^{2}_{P} / 2 (the variance of the mean of two parents  the midparent) as s^{2}_{X}, it can be seen that ß_{MPO} = [½ s^{2}_{A}] / [½ s^{2}_{P}] = h^{2} .^{[48]} Next, utilizing cov(PO) = [ ½ s^{2}_{A} + ½ s^{2}_{D} ] as cov(XY), and s^{2}_{P} as s^{2}_{X}, it is seen that 2 ß_{PO} = [ 2 (½ s^{2}_{A} + ½ s^{2}_{D} )] / s^{2}_{P} = H^{2} .
Analysis of epistasis has previously been attempted via an interaction variance approach of the type s^{2}_{AA} , and s^{2}_{AD} and also s^{2}_{DD}. This has been integrated with these present covariances in an effort to provide estimators for the epistasis variances. However, the findings of epigenetics suggest that this may not be an appropriate way to define epistasis.
Siblings covariances
Covariance between halfsibs (HS) is defined easily using allelesubstitution methods; but, once again, the dominance contribution has historically been omitted. However, as with the midparent/offspring covariance, the covariance between fullsibs (FS) requires a "parentcombination" approach, thereby necessitating the use of the genemodel correctedcrossproduct method; and the dominance contribution has not historically been overlooked. The superiority of the genemodel derivations is as evident here as it was for the Genotypic variances.
Halfsibs of the same commonparent (HS)
The sum of the crossproducts { commonparent frequency * halfbreedingvalue of one halfsib * halfbreedingvalue of any other halfsib in that same commonparentgroup } immediately provides one of the required covariances, because the effects used [breeding values  representing the allelesubstitution expectations] are already defined as deviates from the genotypic mean [see section on "Allele substitution  Expectations and deviations"]. After simplification. this becomes: cov(HS)_{A} = ½ pq a^{2} = ¼ s^{2}_{A} .^{[11]} ^{:132–141}^{[12]} ^{:134–147} However, the substitution deviations also exist, defining the sum of the crossproducts { commonparent frequency * halfsubstitutiondeviation of one halfsib * halfsubstitutiondeviation of any other halfsib in that same commonparentgroup }, which ultimately leads to: cov(HS)_{D} = p^{2} q^{2} d^{2} = ¼ s^{2}_{D} . Adding the two components gives:
cov(HS) = cov(HS)_{A} + cov(HS)_{D} = ¼ s^{2}_{A} + ¼ s^{2}_{D} .
Fullsibs (FS)
As explained in the introduction, a method similar to that used for midparent/progeny covariance is used. Therefore, an unadjusted sum of crossproducts (USCP) using all products—{ parentpairfrequency * the square of the offspringgenotypemean }—is adjusted by subtracting the {overall genotypic mean}^{2} as correction factor (CF). In this case, multiplying out all combinations, carefully gathering terms, simplifying, factoring, and cancellingout is very protracted. It eventually becomes:
cov(FS) = pq a^{2} + p^{2} q^{2} d^{2} = ½ s^{2}_{A} + ¼ s^{2}_{D} , with no dominance having been overlooked.^{[11]} ^{:132–141}^{[12]} ^{:134–147}
Applications (siblings)
The most useful application here for genetical statistics is the correlation between halfsibs. Recall that the correlation coefficient (r) is the ratio of the covariance to the variance [see section on "Associated attributes" for example]. Therefore, r_{HS} = cov(HS) / s^{2}_{all HS together} = [¼ s^{2}_{A} + ¼ s^{2}_{D} ] / s^{2}_{P} = ¼ H^{2} .^{[49]} The correlation between fullsibs is of little utility, being r_{FS} = cov(FS) / s^{2}_{all FS together} = [½ s^{2}_{A} + ¼ s^{2}_{D} ] / s^{2}_{P} . The suggestion that it "approximates" (½ h^{2}) is poor advice.
Of course, the correlations between siblings are of intrinsic interest in their own right, quite apart from any utility they may have for estimating heritabilities or genotypic variances.
It may be worth noting that [ cov(FS)  cov(HS)] = ¼ s^{2}_{A} . Experiments consisting of FS and HS families could utilize this by using intraclass correlation to equate experiment variance components to these covariances [see section on "Coefficient of relationship as an intraclass correlation" for the rationale behind this].
The earlier comments regarding epistasis apply again here [see section on "Applications (Parentoffspring"].
Selection
Basic principles
Selection operates on the attribute (phenotype), such that individuals that equal or exceed a selection threshold (z_{P})' become effective parents for the next generation. The proportion they represent of the base population is the selection pressure. The smaller the proportion, the stronger the pressure. The mean of the selected group (P_{s}) is superior to the basepopulation mean (P_{0}) by the difference called the selection differential (S). All these quantities are phenotypic. To "link" to the underlying genes, a heritability (h^{2}) is used, fulfilling the role of a coefficient of determination in the biometrical sense. The expected genetical change—still expressed in phenotypic units of measurement—is called the genetic advance (?G), and is obtained by the product of the selection differential (S) and its coefficient of determination (h^{2}). The expected mean of the progeny (P_{1}) is found by adding the genetic advance (?G) to the base mean (P_{0}). The graphs to the right show how the (initial) genetic advance is greater with stronger selection pressure (smaller probability). They also show how progress from successive cycles of selection (even at the same selection pressure) steadily declines, because the Phenotypic variance and the Heritability are being diminished by the selection itself. This is discussed further shortly.
Thus ?G = S h^{2}.^{[12]} ^{:1710181}
and P_{1} = P_{0} + ?G .^{[12]} ^{:1710181}
The narrowsense heritability (h^{2}) is usually used, thereby linking to the additive genetic variance (s^{2}_{A}) . However, if appropriate, use of the broadsense heritability (H^{2}) would connect to the genotypic variance (s^{2}_{G}) ; and even possibly an allelic heritability [ h^{2} = (s^{2}_{a}) / (s^{2}_{P}) ] might be contemplated, connecting to (s^{2}_{a} ).
To apply these concepts before selection actually takes place, and so predict the outcome of alternatives (such as choice of selection threshold, for example), these phenotypic statistics are reconsidered against the properties of the Normal Distribution, especially those concerning truncation of the superior tail of the Distribution. In such consideration, the standardized selection differential (i) and the standardized selection threshold (z) are used instead of the previous "phenotypic" versions. The phenotypic standard deviate (s_{P(0)})' is also needed. This is described in a subsequent section.
Therefore, ?G = (i s_{P}) h^{2}, .....where (i s_{P(0)}) = S previously.^{[12]} ^{:1710181}
The text above noted that successive ?G declines because the "input" [the phenotypic variance ( s^{2}_{P} )] is reduced by the previous selection.^{[12]}^{:1710181} The heritability also is reduced. The graphs to the left show these declines over ten cycles of repeated selection during which the same selection pressure is asserted. The accumulated genetic advance (S?G) has virtually reached its asymptote by generation 6 in this example. This reduction depends partly upon truncation properties of the Normal Distribution, and partly upon the heritability together with meiosis determination ( b^{2} ). The last two items quantify the extent to which the truncation is "offset" by new variation arising from segregation and assortment during meiosis.^{[12]} ^{:1710181}^{[25]} This is discussed soon, but here note the simplified result for undispersed random fertilization (f = 0).
Thus : s^{2}_{P(1)} = s^{2}_{P(0)} [1  i ( iz) ½ h^{2}], where i ( iz) = K = truncation coefficient and ½ h^{2} = R = reproduction coefficient^{[12]}^{:1710181}^{[25]} This can be written also as s^{2}_{P(1)} = s^{2}_{P(0)} [1  K R ], which facilitates more detailed analysis of selection problems.
Here, i and z have already been defined, ½ is the meiosis determination (b^{2}) for f=0, and the remaining symbol is the heritability. These are discussed further in following sections. Also notice that, more generally, R = b^{2} h^{2}. If the general meiosis determination ( b^{2} ) is used, the results of prior inbreeding can be incorporated into the selection. The phenotypic variance equation then becomes:
s^{2}_{P(1)} = s^{2}_{P(0)} [1  i ( iz) b^{2} h^{2}].
The Phenotypic variance truncated by the selected group ( s^{2}_{P(S)} ) is simply s^{2}_{P(0)} [1  K], and its contained genic variance is (h^{2}_{0} s^{2}_{P(S)} ). Assuming that selection has not altered the environmental variance, the genic variance for the progeny can be approximated by s^{2}_{A(1)} = ( s^{2}_{P(1)}  s^{2}_{E}) . From this, h^{2}_{1} = ( s^{2}_{A(1)} / s^{2}_{P(1)} ). Similar estimates could be made for s^{2}_{G(1)} and H^{2}_{1} , or for s^{2}_{a(1)} and h^{2}_{1} if required.
Alternative ?G
The following rearrangement is useful for considering selection on multiple attributes (characters). It starts by expanding the heritability into its variance components. ?G = i s_{P} ( s^{2}_{A} / s^{2}_{P} ) . The s_{P} and s^{2}_{P} partially cancel, leaving a solo s_{P}. Next, the s^{2}_{A} inside the heritability can be expanded as (s_{A} * s_{A}), which leads to :
?G = i s_{A} ( s_{A} / s_{P} ) = i s_{A} h . Corresponding rearrangements could be made using the alternative heritabilities, giving ?G = i s_{G} H or ?G = i s_{a} h.
Background
Standardized selection  the normal distribution
The entire base population is outlined by the normal curve^{[47]}^{:78–89} to the right. Along the Z axis is every value of the attribute from least to greatest, and the height from this axis to the curve itself is the frequency of the value at the axis below. The equation for finding these frequencies for the "normal" curve (the curve of "common experience") is given in the ellipse. Notice it includes the mean (µ) and the variance (s^{2}). Moving infinitesimally along the zaxis, the frequencies of neighbouring values can be "stacked" beside the previous, thereby accumulating an area that represents the probability of obtaining all values within the stack. [That's integration from calculus.] Selection focuses on such a probability area, being the shadedin one from the selection threshold (z) to the end of the superior tail of the curve. This is the selection pressure. The selected group (the effective parents of the next generation) include all phenotype values from z to the "end" of the tail.^{[50]} The mean of the selected group is µ_{s}, and the difference between it and the base mean (µ) represents the selection differential (S). By taking partial integrations over curvesections of interest, and some rearranging of the algebra, it can be shown that the "selection differential" is S = [ y (s / Prob.)] , where y is the frequency of the value at the "selection threshold" z (the ordinate of z).^{[11]}^{:226–230} Rearranging this relationship gives S / s = y / Prob., the lefthand side of which is, in fact, selection differential divided by standard deviation—that is the standardized selection differential (i). The rightside of the relationship provides an "estimator" for i—the ordinate of the selection threshold divided by the selection pressure. Tables of the Normal Distribution^{[40]} ^{:547–548} can be used, but tabulations of i itself are available also.^{[51]}^{:123–124} The latter reference also gives values of i adjusted for small populations (400 and less),^{[51]}^{:111–122} where "quasiinfinity" cannot be assumed (but was presumed in the "Normal Distribution" outline above). The standardized selection differential (i) is known also as the intensity of selection.^{[12]}^{:174 ; 186}
Finally, a crosslink with the differing terminology in the previous subsection may be useful: µ (here) = "P_{0}" (there), µ_{S} = "P_{S}" and s^{2} = "s^{2}_{P}".
Meiosis determination  reproductive path analysis
The meiosis determination (b^{2}) is the coefficient of determination of meiosis to the process whereby parents generate gametes. Following the principles of standardized partial regression, of which path analysis is a pictoriallyoriented version, Sewall Wright analyzed the paths of geneflow during sexual reproduction, and established the "strengths of contribution" (coefficients of determination) of various components to the overall result.^{[25]}^{[34]} Path analysis includes partial correlations as well as partial regression coefficients (the latter are the path coefficients). Lines with a single arrowhead are directional determinative paths, and lines with double arrowheads are correlation connections. Tracing various routes according to path analysis rules emulates the algebra of standardized partial regression.^{[45]}
The path diagram to the left represents this analysis of sexual reproduction. Of its interesting elements, the important one in the selection context is meiosis. That's where segregation and assortment occur—the processes that partially ameliorate the truncation of the phenotypic variance that arisies from selection. The path coefficients b are the meiosis paths. Those labeled a are the fertilization paths. The correlation between gametes from the same parent (g) is the meiotic correlation. That between parents within the same generation is r_{A}. That between gametes from different parents (f) became known subsequently as the inbreeding coefficient.^{[11]}^{:64} The primes ( ' ) indicate generation (t1), and the unprimed indicate generation t. Here, some important results of the present analysis are given. Sewall Wright interpreted many in terms of inbreeding coefficients.^{[25]}^{[34]}
The meiosis determination (b^{2}) is ½ (1+g) and equals ½ (1 + f_{(t1)}) , implying that g = f_{(t1)}. [Notice that this b^{2} is the coefficient of parentage (f_{AA}) of Pedigree analysis rewritten with a "generation level" instead of an "A" inside the parentheses.] With nondispersed random fertilization, f_{(t1)}) = 0, giving b^{2} = ½, as used in the selection section above. However, being aware of its background, other fertilization patterns can be used as required. Another determination also involves inbreeding  the fertilization determination (a^{2}) equals 1 / [ 2 ( 1 + f_{t} ) ] . Also another correlation is an inbreeding indicator  r_{A} = 2 f_{t} / ( 1 + f_{(t1)} ), also known as the coefficient of relationship (Do not confuse this with the coefficient of kinship—an alternative name for the coancestry coefficient. See introduction to "Relationship" section). This r_{A} reoccurs in the subsection on dispersion and aelection.
These links with inbreeding reveal interesting facets about sexual reproduction that are not immediately apparent. The graphs to the right plot the meiosis and syngamy (fertilization) coefficients of determination against the inbreeding coefficient. There it is revealed that as inbreeding increases, meiosis becomes more important (the coefficient increases), while syngamy becomes less important. The overall role of reproduction [the product of the previous two coefficients  r^{2}] remains the same.^{[52]} This increase in m^{2} is particularly relevant for selection because it means that the selection truncation of the Phenotypic variance is offset to a lesser extent during a sequence of selections when accompanied by inbreeding (which is frequently the case).
Dispersion and selection
Selection basics concern the selection for future parenthood of individuals belonging to a population, on the basis of their phenotypic values. Gamete sampling results in that population becoming dispersed into progeny lines with variable allele and genotype frequencies, and with variable means. This necessitated the partitioning of the genotypic variance into withinline and amongstline components. It remains to superimpose this dispersion structure onto the basics of selection. Sewall Wright's analysis of reproduction (see previous subsection) provides a convenient key to enable this. This key is the coefficient of relationship (r_{A}), which is the genic correlation among individuals in the (t1) generation—that is, it is an intraclass correlation. This is a special class of correlations defined in terms of variance components. In particular, they are generally the ratio of the interclass variance to the total of both interclass and intraclass variances.^{[7]}^{:282–284}^{[40]}^{:294–296} This connection is demonstrated in the following subsection. It is important to affirm this idea because the method of connecting selection and dispersion directly utilizes intraclass correlations.
Coefficient of relationship as an intraclass correlation
It is sometimes "assumed" that the approximations s^{2}_{A(AL)} = 2 f s^{2}_{A(0)} and s^{2}_{A(WL)} = (1 f) s^{2}_{A(0)} and s^{2}_{A(S)} = (1 + f) s^{2}_{A(0)} suffices [see section on dispersion and the genotypic variance], and it is temporally useful here (it is corrected subsequently). Furthermore, recall that the inbreeding coefficient in the tth generation (f_{t}) is the same as the coancestry coefficient (f_{j j’}) in the (t1) generation [see section on Pedigree analysis, et seq]. Putting all of this together,
r_{A(jj’)} = s^{2}_{A(AL)} / s^{2}_{A(S)} = [2f_{jj’} s^{2}_{A(0)}] / [(1 + f_{(t1)}) s^{2}_{A(0)}] = 2f_{jj’} / (1 + f_{(t1)}) (after cancelling the variances) = 2f_{t} / (1 + f_{(t1)}) = Sewall Wright's result.
Thus, the coefficient of relationship is indeed an intraclass correlation in principle.^{[12]} ^{:208–218} The simplifying assumptions now need addressing.
Referring to the section on dispersion and genotypic variance, note that actually s^{2}_{A(AL)} = 2 f s^{2}_{a} and s^{2}_{A(S)} = (1 + f) s^{2}_{A(f)}. After substituting these into the r_{A} equation, and simplifying, the corrected genic intraclass correlation (c_{A}) is:
c_{A} = r_{A} ( a^{2} / a^{2}_{f (t1)} ) .
The difference between the two is important if dominance is non trivial. See the graphs to the right.
Phenotypic intraclass correlation
This is the other prominent intraclass correlation relating selection to dispersion. It is based on partitioning the total phenotypic variance into amongstline and withinline components, which have an underlying genetical cause. Its derivation serves to affirm that a correlation can be constructed with variance components. The following biometrical model forms the skeleton:
X_{ij} = µ + ?_{i} + ?_{ij} ...where... X_{ij} is the phenotype of the jth individual within the ith line, i equals 1 .... g where g = the number of lines, j equals 1 .... n where n = number of individuals within the line, µ is the grand mean, ?_{i} is the line effect of the ith line [the expectation—the deviation between the grand mean and the mean of the individuals within the ith line], and ?_{ij} is the deviation of the jth individual within line i and the mean of all the individuals within line i.^{[7]}^{[47]}
The variance components associated with this model are: s^{2}_{X( i j )} = s^{2}_{?} + s^{2}_{?} = s^{2}_{AL} + s^{2}_{WL} = s^{2}_{P(S)} .
Recalling that the correlation (succinctly stated) is the ratio of the covariance to the variance (or to the geometric mean of two variances if necessary) [see section on "correlated traits"], the covariance [ X( i j ), X( i j’ ) ] needs defining, where j ? j’ being separate individuals within the same line. This covariance can be found as the Expectation of the crossproduct of the modelcomponents defining each individual,^{[47]} as follows:
cov [ X( i j ), X( i j’ ) ] = E{(?_{i} + ? _{ij} ) ( ?_{i} + ? _{ij’} )} , where E is the "Expectation", that is the mean under infinite sampling of all background.^{[7]}^{[47]} Continuing :
cov [ X( i j ), X( i j’ ) ] = E{( ?_{i}) ^{2} } + E{ ?_{i} ? _{ij’} } + E{ ?_{i} ? _{ij} } + E{? _{ij} ? _{ij’} } = s^{2}_{AL} + 0 + 0 + 0 , the first by definition of the variance, and the rest by the fundamental assumption of the independence of effects in the model.^{[7]}^{[47]}
Thus, r_{P} = c_{P} = cov [ X( i j ), X( i j’ ) ] / s^{2}_{X( i j )} = s^{2}_{AL} / [ s^{2}_{AL} + s^{2}_{WL} ] .
This satisfies the biometrical definition of any intraclass correlation, but it ignores the genetical origin of the matter. The very gamete sampling that gave rise to dispersion also has made likely that the uniting gametes were not independent, with the result that the E{? _{ij} ? _{ij’} } ? 0 after all. Sewall Wright's reproductive paths identify two correlations that lead to this lack of independenc: g ( the meiosis correlation—f_{(t1)}) and f (the gamete correlation of separate parents—f_{t}). (Refer to that section). The combined effect of these can be defined as r_{g} = f_{t} + f_{(t1)} , which has a striking resemblance to de novo inbreeding plus carryover inbreeding of previous sections !! Now, define g^{2} as the coefficient of determination that quantifies the genotypic component of s^{2}_{WL} , giving consequently s^{2}_{WL(g)} = g^{2} s^{2}_{WL} . Finally, recalling that covariance equals correlation multiplied by variance, the genotypic covariance amongst individuals within the same dispersion line cov_{WL(g)} = r_{g} s^{2}_{WL(g)} = (f_{t} + f_{(t1)}) [g^{2} s^{2}_{WL}] . That is to say:
E{? _{ij} ? _{ij’} } = r_{g} s^{2}_{WL(g)} ....[instead of 0], giving:
k_{P} = [ s^{2}_{AL} + r_{g} s^{2}_{WL(g)} ] / s^{2}_{P(S)} = r_{P} + [ r_{g} s^{2}_{WL(g)} ] / s^{2}_{P(S)} , after including the effects of inbreeding. All symbols are defined within the derivation. The implications of this are that if s^{2}_{AL} and s^{2}_{WL} were estimated from the Meansquares of an actual analysisofvariance of a "nursery", the "apparent" s^{2}_{AL} would be biased upward by the amount [ r_{g} s^{2}_{WL(g)} ] / s^{2}_{P(S)} , for which adjustment would be needed. Unbiased r_{P} (c_{P}) could then be calculated.
Relating intraclass correlations to dispersion and to Heritabilities
Fundamentals
Both phenotypic and genic intraclass correlations have been shown to be the ratios of their respective amongstline variances to their respective total variances, as defined within the simple dispersion model. Therefore:^{[12]} ^{:208–218}
s^{2}_{P(AL)} = c_{P} s^{2}_{P(S)} , ...and... s^{2}_{P(WL)} = (1  c_{P}) s^{2}_{P(S)} .
Similarly, s^{2}_{A(AL)} = c_{A} s^{2}_{A(S)} , ...and... s^{2}_{A(WL)} = (1  c_{A}) s^{2}_{A(S)} .
These are without error, provided unbiased intraclass correlations (c_{P} and c_{A}) are used [see the sections above]. The amongstline phenotypic variance is the variance of dispersed progeny means arising from genetic drift and/or gamete relationship. It is therefore a "genotypic variance" (of dispersed progenies). The withinline phenotypic variance is the variance of individuals within progeny lines, being their "genotypic variance" confounded probably with their "environmental variance". The "genotypic" component of this was [g^{2} s^{2}_{P(WL)} ] in the amendment giving rise to k_{P} in the section above. The genic variances are the previouslydiscussed dispersion genic variances.
It is also possible to define genotypic (c_{G}) and dominance (c_{D}) intraclass correlations that can be used in parallel ways to relate to their respective amongstline and withinline variance components.
Corollaries
Recalling the comment above that s^{2}_{P(AL)} = s^{2}_{G(AL)} , it is often therefore equated to s^{2}_{A(AL)} = c_{A} s^{2}_{A(S)} . Now, recalling that r_{P} = s^{2}_{P(AL)} / s^{2}_{P(S)} , this new substitution leads to r_{P} = c_{A} s^{2}_{A(S)} / s^{2}_{P(S)}. Further, recall that h^{2}_{Individuals} = s^{2}_{A(S)} / s^{2}_{P(S)} . It can thus be seen that r_{P} = c_{A} h^{2} _{Indiv.} . That is, it is an "overview heritability" of line dispersion ! [Notice that this equality applies to r_{P} : not to k_{P}.] A simple "withinline" heritability { [(1  c_{A}) / (1  c_{P})] h^{2} _{Indiv.} }, and "amongstline" heritability { [ c_{A} / c_{P}] h^{2} _{Indiv.} } can also be constructed, but more utilitarian versions follow.^{[12]} ^{:208–218} Notice that this simple "withinline" heritability is not the same as g^{2} of the previous section on the "phenotypic intraclass correlation", being a "narrowsense" heritability instead of a "broadsense" one. The g^{2} would be [(1  c_{G}) / (1  c_{P})] H^{2} _{Indiv.} instead.
Amongstline selection
Within a natural dispersed bulk of progeny lines, it is probably impossible to identify the lines and their respective members: but, within the "nursery" of a controlled selection programme, parents and offspring are managed beforehand, and both progenylines and their members are certainly identifiable. It is thus possible to analyse the nursery as an "experiment", and to conduct a simple "single hypothesis" (oneway) analysisofvariance (ANOVA) on it, extracting estimates of means and variancecomponents together with their standard errors.^{[7]}^{[40]}^{[47]} If this was possible for the natural bulk, this ANOVA approach could also be used.
As a result, amongstline selection can actually be effected: "observe" the line means and apply (to these) amongstline selection as derived from the selection basics. Because observed means are usable, the variance of an estimated mean has to be added to the s^{2}_{(AL)}, whether it be phenotypic or genic. In general, the variance of an estimated mean s^{2}_{?} = s^{2}_{y} / n_{?} .^{[7]}^{[40]}^{[47]} Similarly, the amongstline "narrowsense" heritability has to be redefined in this new light. Finally, all these modifications have to be amalgamated into one compoundcoefficient (?_{AL'} ) so as to convert at once the ?G_{indiv} into ?G_{AL'} . The one remaining modification is to realize that selection pressure (Prob) now applies to the proportion of means that are selected, and does not focus on the proportion of individuals. Consequently, the selection threshold (z), its ordinate (y), and the intensity of selection (i) are all focused in this way.
Therefore,^{[12]}^{:208–218} s^{2}_{P(AL' )} = s^{2}_{P(AL)} + s^{2}_{P( ? )} = s^{2}_{P(AL)} + ( s^{2}_{P(WL)} / n_{WL} ) . That is: s^{2}_{P(AL' )} = c_{P} s^{2}_{P(S)} + [ (1c_{P}) s^{2}_{P(S)} ] / n_{WL} .
After gathering of terms, and simplifying, this becomes: s^{2}_{P(AL' )} = s^{2}_{P(S)} .
In a like manner, s^{2}_{A(AL' )} = s^{2}_{A(S)} .
The ratio of the latter to the former furnishes the appropriate "narrowsense" heritability, which, after simplifying, becomes: h^{2}_{AL'} = h^{2}_{Indiv} { [c_{A} (n_{WL}1) + 1] / [c_{P} (n_{WL}1) + 1] } .
Lastly, ?G_{AL'} = i s_{P(AL' )} h^{2}_{AL'} .... = .... [ i s_{P(S)} h^{2}_{Indiv} ] ^{(1/2)} , where all symbols have been defined previously.
Finally, ?G_{Cbnd} = [ i s_{P(S)} h^{2}_{Indiv} ] ?_{Cbtn} = ?G_{Indiv} ?_{Cbtn}, remembering to use the appropriate i as discussed above.
Relative efficiencies of selection strategies
Comparison of the values of the various ?G conversion coefficients (?_{x}) provides an immediate measure of the relative merits of the four strategies over time (generations) under various inbreeding regimes (dispersion strengths). For this purpose, the ?_{Indiv} can be set at 1. Using the ?_{x} criteria, combined selection is found to be the best under every inbreeding regime. Therefore, the simplest method to visualize the relative efficiencies is to obtain the ratios of each ?_{x (t)} to ?_{Cbnd (t)}. Graphs to the right show these ?ratios over ten cycles of selection for successive halfsib inbreeding [see section on halfsib crossing]. It is immediately apparent that, for this inbreeding regime, withinline selection has no value for making genetic advance (?G). Its only purpose might be for purification of breedingstocks.
Application note: Even though combined selection is most efficient for ?G, it is a twostage selection activity, and therefore is more costly in time and money. It may not be so highly desirable when these new criteria are considered. For that reason, the graphs include a "0.9 efficiency judgment" line as well as the ?ratios themselves. This could be used in the following way. If a plant breeder was executing a "line selection" programme under such pollen management,^{[32]} he might decide to utilize cheaper "onepass" selection strategies that fell well above the 0.9 cutoff line in lieu of combined selection. On this basis, using these graphs, he might choose to use overall individual selection (that is "ignore" his dispersion structure in the nursery) for the first two cycles, then use combined selection for four further cycles, and finish with four cycles of amongstline selection. Of course, this is only one possibility. He would not "abandon" his dispersion maintenance, however, and would continue with the rigorous pollen management the halfsib crossing demands.
Genetic drift and selection
The previous sections treated dispersion as an "assistant" to selection, and it became apparent that the two work well together. In quantitative genetics, selection is usually examined in this "biometrical" fashion, but the changes in the means (as monitored by ?G) reflect the changes in allele and genotype frequencies beneath this surface. Referral to the section on "Genetic drift" brings to mind that it also effects changes in allele and genotype frequencies, and associated means; and that this is the companion aspect to the dispersion considered here ("the other side of the same coin"). However, these two forces of frequency change are seldom in concert, and may often act contrary to each other. One (selection) is "directional" being driven by selection pressure acting on the phenotype: the other (genetic drift) is driven by "chance" at fertilization (binomial probabilities of gamete samples). If the two tend towards the same allele frequency, their "coincidence" is the probability of obtaining that frequencies sample in the genetic drift: the likelihood of their being "in conflict", however, is the sum of probabilities of all the alternative frequency samples. In extreme cases, a single syngamy sampling can undo what selection has achieved, and the probabilities of it happening are available. It is important to keep this in mind. However, genetic drift resulting in sample frequencies similar to those of the selection target does not lead to so drastic an outcome—instead slowing progress towards selection goals.
Although some genes have only an effect on a single trait, many genes have an effect on various traits, which is called pleiotropy. Because of this, a change in a single gene effects all those traits. This is calculated using covariances, and the phenotypic covariance (cov_{P}) between two traits can be partitioned in the same way as the variances described above [e.g., genic (cov_{A} ), dominance (cov_{D}), and environment (cov_{E} )] . In general, the correlation coefficient is the ratio of the covariance to the geometric mean of the two variances of the traits.^{[47]} ^{:196–198} Various correlation coefficients can be obtained, using the appropriate partitions of variances and covariances. Of course, the Phenotypic correlation is the "usual" correlation of Statistics/Biometrics.
 \mbox{Phenotypic correlation} = \frac{\mathrm{cov}(P_{1}, P_{2})}{\sqrt}} ...and... \mbox{Genotypic correlation} = \frac{\mathrm{cov}(G_{1}, G_{2})}{\sqrt}} ...and also... \mbox{Environmental correlation} = \frac{\mathrm{cov}(E_{1}, E_{2})}{\sqrt}} .
The genic correlation (genetic correlation) is of particular interest, especially in quantifying the correlated effects of selection. It is as follows:
 \mbox{Genic correlation} = \frac{\mathrm{cov}(A_{1}, A_{2})}{\sqrt}}
See also
 Artificial selection
 Diallel cross
 Douglas Scott Falconer
 Ewens's sampling formula
 Experimental evolution
 Genetic architecture
 Genetic distance
 Heritability
 Ronald Fisher
Footnotes and references
 ^ Anderberg, Michael R. (1973). Cluster analysis for applications. New York: Academic Press.
 ^ Mendel, Gregor (1866). "Versuche über Pflanzen Hybriden". Verhandlungen naturforschender Verein in Brünn iv.
 ^ ^{a} ^{b} ^{c} Mendel, Gregor; Bateson, William [translator] (1891). "Experiments in plant hybridisation". J. Roy. hort. Soc. (London) xxv: 54–78.
 ^ The Mendel G.; Bateson W. (1891) paper, with additional comments by Bateson, is reprinted in: Sinnott E.W.; Dunn L.C.; Dobzhansky T. (1958). "Principles of genetics"; New York, McGrawHill: 419443. Footnote 3, page 422 identifies Bateson as the original translator, and provides the reference for that translation.
 ^ A QTL is a region in the DNA genome that affects, or is associated with, quantitative phenotypic traits.
 ^ ^{a} ^{b} Fisher, R. A. (1918). "The correlation between relatives on the supposition of Mendelian inheritance.". Trans. Roy. Soc, (Edinburgh) 52: 399–433.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} ^{g} ^{h} ^{i} ^{j} ^{k} ^{l} ^{m} Steel, R. G. D.; Torrie, J. H. (1980). Principles and procedures of statistics. (2 ed.). New York: McGrawHill.
 ^ Other symbols are sometimes used, but these are common.
 ^ The allele effect is the average phenotypic deviation of the homozygote from the midpoint of the two contrasting homozygote phenotypes at one locus, when observed over the infinity of all background genotypes and environments. In practice, estimates from large unbiased samples substitute for the parameter.
 ^ The dominance effect is the average phenotypic deviation of the heterozygote from the midpoint of the two homozygotes at one locus, when observed over the infinity of all background genotypes and environments. In practice, estimates from large unbiased samples substitute for the parameter.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} ^{g} ^{h} ^{i} ^{j} ^{k} ^{l} ^{m} ^{n} ^{o} ^{p} ^{q} ^{r} ^{s} ^{t} ^{u} ^{v} ^{w} ^{x} ^{y} ^{z} ^{aa} ^{ab} ^{ac} ^{ad} Crow, J. F.; Kimura, M. (1970). An introduction to population genetics theory. New York: Harper & Row.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} ^{g} ^{h} ^{i} ^{j} ^{k} ^{l} ^{m} ^{n} ^{o} ^{p} ^{q} ^{r} ^{s} ^{t} ^{u} ^{v} ^{w} ^{x} ^{y} ^{z} ^{aa} ^{ab} ^{ac} ^{ad} ^{ae} ^{af} ^{ag} ^{ah} ^{ai} ^{aj} ^{ak} ^{al} ^{am} ^{an} ^{ao} ^{ap} ^{aq} Falconer, D. S.; Mackay, Trudy F. C. (1996). Introduction to quantitative genetics (Fourth ed.). Harlow: Longman.
 ^ Mendel commented on this particular tendency for F1 > P1, i.e., evidence of hybrid vigour in stem length. However, the difference may not be significant. (The relationship between the range and the standard deviation is known [Steel and Torrie (1980): 576], permitting an approximate significance test to be made for this present difference.)
 ^ Richards, A. J. (1986). Plant breeding systems. Boston: George Allen & Unwin.
 ^ Jane Goodall Institute. "Social structure of chimpanzees.". Chimp Central. Retrieved 20 August 2014.
 ^ Gordon, Ian L. (2000). "Quantitative genetics of allogamous F2: an origin of randomly fertiliized populations.". Heredity 85: 43–52.
 ^ An F2 derived by self fertilizing F1 individuals (an autogamous F2), however, is not an origin of a randomly fertilized population structure. See Gordon (2001).
 ^ Castle, W. E. (1903). "The law of heredity of Galton and Mendel and some laws governing race improvement by selection.". Proc. Amer. Acad, Sci. 39: 233–242.
 ^ Hardy, G. H. (1908). "Mendelian proportions in a mixed population.". Science 28 (706): 49–50.
 ^ Weinberg, W. (1908). "Über den Nachweis der Verebung beim Menschen.". Jahresh. Verein f. vaterl. Naturk, Württem. 64: 368–382.
 ^ Usually in science ethics, a discovery is named after the earliest person to propose it. Castle, however, seems to have been overlooked: and later when refound, the title "Hardy Weinberg" was so ubiquitous it seemed too late to update it. Perhaps the "Castle Hardy Weinberg" equlilbrium would be a good compromise?
 ^ ^{a} ^{b} Gordon, Ian L. (1999). "Quantitative genetics of intraspecies hybrids.". Heredity 83: 757–764.
 ^ Gordon, Ian L. (2001). "Quantitative genetics of autogamous F2.". Hereditas 134 (3): 255–262.
 ^ Wright, S. (1917). "The average correlation within subgroups of a population.". J. Wash. Acad. Sci. 7: 532–535.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} ^{g} Wright, S. (1921). "Systems of mating. I. The biometric relations between parent and offspring.". Genetics 6: 111–123.
 ^ Sinnott, Edmund W.; Dunn, L. C.; Dobzhansky, Theodosius (1958). Principles of genetics. New York: McGrawHill.
 ^ ^{a} ^{b} ^{c} Fisher, R. A. (1999). The genetical theory of natural selection. ("variorum" ed.). Oxford: Oxford University Press.
 ^ ^{a} ^{b} Cochran, William G. (1977). Sampling techniques. (Third ed.). New York: John Wiley & Sons.
 ^ This is outlined subsequently in the genotypic variances section.
 ^ Both are used commonly.
 ^ See the earlier citations.
 ^ ^{a} ^{b} Allard, R. W. (1960). Principles of plant breeding. New York: John Wiley & Sons.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} ^{g} Gordon, I.L. (2003). "Refinements to the partitioning of the inbred genotypic variance". Heredity 91 (1): 85–89.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} Wright, Sewall (1951). "The genetical structure of populations.". Annals of Eugenics 15: 323–354.
 ^ ^{a} ^{b} ^{c} Mather, Kenneth; Jinks, John L. (1971). Biometrical genetics (2 ed.). London: Chapman & Hall.
 ^ These have been translated from Mather's symbols into Fisherian ones to facilitate the comparison.
 ^ Covariance is the covariability between two sets of data—in this case the a and the d. Similarly to the variance, it is based on a sum of crossproducts (SCP) instead of a SS. From this, it is clear therefore that the variance is but a special form of the covariance.
 ^ Hayman, B. I. (1960). "The theory and analysis of the diallel cross. III.". Genetics 45: 155–172.
 ^ It has been observed that when p = q, or when d = 0, a [= a+(qp)d] "reduces" to a. In such circumstances, s^{2}_{A} = s^{2}_{a}  but only numerically. They still have not become the one and the same identity.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} ^{g} Snedecor, George W.; Cochran, William G. (1967). Statistical methods. (Sixth ed.). Ames: Iowa State University Press.
 ^ ^{a} ^{b} ^{c} Kendall, M. G.; Stuart, A. (1958). The advanced theory of statistics. Volume 1. (2nd ed.). London: Charles Griffin.
 ^ It is common practice not to have a subscript on the experimental "error" variance.
 ^ This type of varianceratio is an example of a coefficient of determination. It is used particularly in regression analysis. A standardized version of regression analysis is path analysis. Here, standardizing means that the data were first divided by their own experimental standard errors to unify the scales for all attributes.
 ^ Dohm, M. R. (2002). "Repeatability estimates do not always set an upper limit to heritibility.". Functional ecology 16: 273–280.
 ^ ^{a} ^{b} ^{c} ^{d} Li, Ching Chun (1977). Path analysis  a Primer (Second printing with Corrections ed.). Pacific Grove: Boxwood Press.
 ^ ^{a} ^{b} Draper, Norman R.; Smith, Harry (1981). Applied regression analysis. (Second ed.). New York: John Wiley & Sons.
 ^ ^{a} ^{b} ^{c} ^{d} ^{e} ^{f} ^{g} ^{h} ^{i} ^{j} Balaam, L. N. (1972). Fundamentals of biometry. London: George Allen & Unwin.
 ^ In the past, both forms of parentoffspring covariance have been applied to this task of estimating h^{2}, but, as noted in the subsection above, only one of them (cov(MPO)) is actually appropriate. The cov(PO) is useful, however, for estimating H^{2} as seen in the main text following.
 ^ Note that texts that ignore the dominance component of cov(HS) erroneously suggest that r_{HS} "approximates" ( ¼ h^{2} ).
 ^ Theoretically, the tail is infinite, but in practice there is a quasiend.
 ^ ^{a} ^{b} Becker, Walter A. (1967). Manual of procedures in quantitative genetics. (Second ed.). Pullman: Washington State University.
 ^ There is a small "wobble" arising from the fact that b^{2} alters one generation behind a^{2}  examine their inbreeding equations.
Further reading
 Lynch M & Walsh B (1998). Genetics and Analysis of Quantitative Traits. Sinauer, Sunderland, MA.
 Roff DA (1997). Evolutionary Quantitative Genetics. Chapman & Hall, New York.
 Seykora, Tony. Animal Science 3221 Animal Breeding. Tech. Minneapolis: University of Minnesota, 2011. Print.
External links
 The Breeder's Equation
 Quantitative Genetics Resources by Michael Lynch and Bruce Walsh, including the two volumes of their textbook, Genetics and Analysis of Quantitative Traits and Evolution and Selection of Quantitative Traits.
 Resources by Nick Barton et al. from the textbook, Evolution.
 The GMatrix Online

