Kinship
The inclusion of research subjects that vary in their levels of relatedness can inflate the proportion of phenotypic variance in any trait under study that is due to the genetic component of that variance. This inflation of phenotypic variance obscures treatment effects so that a greater number of research subjects is required to demonstrate a statistically significant treatment effect. If identical twins or subjects matched for genetic characteristics that influence phenotypic traits of interest (such as MHC and ABO matched donors and recipients in transplantation research), are not available, the selection of unrelated subjects is advisable.
The level of kinship between any two pairs of animals can be estimated by the parameter "r." The expected values of "r" are 1 for identical twins, 0.5 for full sibs and fraternal twins, 0.25 for half sibs, uncle/niece or uncle/nephew or aunt/nephew or aunt/niece, 0.125 for first cousins and 0 for unrelated animals.
The value of "r" can be estimated by several different methods that have been developed using either di-allelic markers (those with only two alleles), such as SNPs, or multiple allelic markers, such as microsatellite (STR) loci. While multiple allelic markers are more effective than the same number of di-allelic markers, equal efficiency (level of accuracy) can be achieved with di-allelic markers by using a much larger number of such markers.
Tests done by the Molecular Anthropology Laboratory [MAL] use up to 24 STRs or up to 96 SNPs to estimate "r." The laboratory"s method of estimating "r" is that developed by Lynch (1988) as modified by Li et al. (1992) for use with DNA "fingerprints" because it is simultaneously the simplest to estimate and provides among the lowest standard errors in the estimate of "r." By this method the coefficient of relationship between animals X and Y based on a single locus is estimated by:
The value of "S_(X Y)" = 1 when animals X and Y share identical genotypes, 0.75 when X is homozygous for an allele shared with Y when Y is heterozygous, 0.5 when both X and Y are heterozygous and share only one allele in common and 0 when X and Y share no alleles in common.
The kinship coefficient becomes the average of the "r_(X Y)"values over all N (24 STR or 96 SNP) loci with a standard error (SE) for unrelated pairs of approximately 1/√N and √1/N(n-1) where N and n are respectively, the number of loci used and the average number of alleles at each locus (Lynch and Ritland, 1999) for di-allelic and multiple-allelic loci, respectively.
The value of n for the 24 STRs in the laboratory"s panels is approximately 7 and 10 for Indian and Chinese rhesus macaques and between 5 and 9 for the different regional populations of longtail macaques. Thus, the SE for estimating "r" is approximately 0.1 and 0.04, respectively for unrelated animals. While estimated values of "r_(X Y)" for true relatives more distantly related than full first cousins (r=0.125) cannot be easily distinguished from that for unrelated pairs of animals, a limited number of animals less closely related than first cousins in a control or experimental group will have minimal influence on the genotypic component of the phenotypic variance.
The Molecular Anthropology Laboratory will report to you the value of "r" for all possible pairs of samples submitted as well as the average value over all pairs. The subjects selected for inclusion in a given research project should share the lowest possible values of "r_(X Y)". An accurate estimate of "S_(O)" requires that you know the country of origin of the animals being tested.
Li, C. C., D. E. Weeks and A. Chakravarti, 1993 Similarity of DNA assume that at least 10 or so loci are sampled, because fingerprints due to chance and relatedness. Hum. Hered. 43: with fewer loci, the tradeoff involving r favors more loci 45-52.
Lynch, M., 1988 Estimation of relatedness by DNA fingerprinting. Mol. Biol. Evol. 5: 584-599.
Lynch,M., and K. Ritland, 1999 Estimation of pairwise relatedness sample sizes in both accuracy and precision. Theupward with molecular markers. Genetics 152: 1753-1766.
