Graphical depiction of the SNP-string method presented here (see text).

<p>Subject-vectors (strings of 0s, 1s, and 2s) are searched for homozygous and single-site heterozygus individuals (A). These “unambiguous” combinations are decomposed into their constituent “unambiguous” SNP-string combinations and then a list of all the unambiguous SNP-strings that have been “identified” is compiled (A). Following this, the entire subject-vector list is decomposed into the possible combination categories (B). The “indeterminate (unique)” decompositions are used to “identify” additional SNP-strings, which are then added to the “identified” SNP-string list (B and C). The entire decomposition is then repeated until either no further SNP-strings can be added to the “identified” list or more than 4 decompositions have taken place. The final list are considered “identified” SNP-strings (C and D). Unique decompositions (of “identified” SNP- strings) and unambiguous combinations are used to calculate the allellic frequencies used for resolving the “conflicted” decompositions (E). Following this, a list of all possible addtional “novel” SNP-strings is compiled from the “indeterminate (conflicted)” decompositions (F). This list is then combined with the “identified” list to make the “complete” SNP-string list (F) and the decomposition repeated. By definition, using this “complete” list, every subject-vector will be either an unambiguous combination or a unique or conflicted decomposition. Allellic frequencies are recalculated from the unique decompositions and unambiguous combinations and these frequencies used to resolve the conflicted decompositions. Those SNP-strings that never selected (i.e., have a zero final frequency) are then dropped from the “complete” list and the decomposition repeated until all SNP-strings on the “complete” list have a non-zero final frequency (G). In the final step, persons with uncommon alleles (selected in less than 9 individuals) are then reassessed for alternative possible combinaitions with “novel” SNP-strings, these novel SNP-strings are added to the “complete” list, and the process described above is again repeated until all SNP-strings on the “complete” list have a non-zero final frequency and the SNP-string composition of every subject-vector has been selected.