Sample-distance

Sample-Distance and Similarity Matrices

Assessment of similarity between studied individuals is useful to identify sample duplications or unexpected outliers. It is critical to adjust for hidden population structure. A pairwise similarity can be calculated by counting the alleles descended from a common ancestor (i.e. identity by descent; IBD) and/or alleles that are the same, irrespective of whether they are inherited from a recent ancestor (i.e. identity by state; IBS).

Pedigree information allows us to figure out the expected proportion of shared alleles by descent and thus IBD distances within families. However, it is not uncommon for the proposed pedigree to be wrong. With genotyping arrays and sequencing, marker information can be used to calculate IBS or IBD similarity matrices and predict the actual pedigree structure.

VanRaden (2008) and Yang et al. (2010) described the genomic relationship matrix (GRM) as a a covariance matrix which uses identical by state (IBS) information scaled by the allele frequencies, as shared rare alleles are more likely to be IBD than common alleles. These methods do not explicitly differentiate between IBD and IBS information Clark et al. (2012).

GRM uses the minor allele frequencies (MAFs) of genotyped SNPs assuming a homogeneous population structure which is not always a true assumption. KING-robust estimation is an alternative approach that does not need MAFs. This technique is more trusted in mixed-population datasets (unless the parents are from very different populations).

In the GRLS study, use either technique can be used safely. Therefore, the KING-robust kinship estimator will be used in this tutorial. According to the PLINK2 documentation, the KING kinship coefficients are scaled such that duplicate samples have a kinship of 0.5. Accordingly, first-degree relations (parent-child, full siblings) correspond to ~0.25, second-degree relations correspond to ~0.125, etc. It is conventional to use a cutoff of ~0.354 (the geometric mean of 0.5 and 0.25) to screen for monozygotic twins and duplicate samples, ~0.177 to add first-degree relations, etc.