A genetic value analysis ranks animals according to their genetic value to the breeding colony, based on an algorithm that incorporates measures of mean kinship to the rest of the colony, as well as the probability that an animal contains rare founder alleles that should be conserved in the colony.

Genome uniqueness is an estimate produced by a gene-drop simulation. The analysis drops alleles down the pedigree many times (the number of iterations you choose) and reports the average result, so each genome uniqueness value comes with a sampling standard error -- the +/- shown in the guSE column, and summarized as the largest value on the Summary tab. Running more iterations makes that +/- smaller (it shrinks roughly in proportion to one over the square root of the number of iterations), so there is no single iteration count that is right for every colony: how many you need depends on your pedigree.

Founder genome equivalents (the FG value on the Summary tab) is likewise an estimate from the same gene-drop simulation, so it too is reported with a sampling standard error -- the +/- shown next to FG. As with genome uniqueness, running more iterations makes that +/- smaller (it shrinks roughly in proportion to one over the square root of the number of iterations). At a small number of iterations the FG estimate also carries a slight bias that the +/- does not capture, so prefer more iterations when the precision of FG matters.

Gene diversity and effective population size. The Summary tab also reports gene diversity (GD) and two effective population size estimates -- a demographic sex-ratio effective size and a variance effective size, both over the current living breeders. These, with their formulas and idealizing assumptions, are defined in the Population Genetics Terms reference panel on the Summary Statistics tab.

Keep two ideas separate. The +/- tells you how precise the genome uniqueness number is. What actually decides which animals are chosen for breeding is the order in which animals are ranked. A small +/- does not by itself mean the ranking has settled. The package provides a tool for exactly this check -- the gvaConvergence() function recomputes the ranking across a range of iteration counts and reports the smallest number of gene-drop iterations at which the selection order for your pedigree stops changing.

For a full description of the algorithm, see the Genetic Value Analysis and Breeding Group Description tab.

Kinship overrides. If you upload outside-information kinship values on this tab, they replace the pedigree-derived kinship for the listed pairs and feed the rankings, breeding groups, and summary statistics regardless of tab order. They change the kinship value only: in the Summary Statistics relationship table the relationship label stays pedigree-derived, so a label and its overridden value can disagree. The iteration-convergence check above (gvaConvergence()) applies these overrides as well, ranking on the overridden kinship. For an animal missing one parent, an override only refines that pair's kinship value; the unknown-parent correction is kept for every such animal. Mean kinship for an animal missing a parent is an estimate that tends to underestimate relatedness, because the unknown parent's relatives are unrecorded (Vinson & Raboin 2015).