Optimizing the information increase through the addition of relatives and genetic markers in identification and kinship cases.

Highlights • In most pairwise kinship problems weak statistics result from genetic analysis. • Statistics improve when a third individual or a larger set of markers are analysed. • We identify the relative whose analysis will provide greater genetic information. • We recommend the minimal number of markers for each type of kinship problem. • We provide an informatics tool to calculate false discovery rates. Abstract In kinship testing powerful statistical results are usually obtained when genetic information is required to be shared between a set of individuals, under the assumption of one of the hypotheses. This is the case when the hypotheses parenthood or identity are compared with the hypothesis of no relatedness, even when only a pair of individuals is analyzed. In this work we will consider kinship problems where the sharing of genetic information is not required when a pair of individuals is analyzed, such as is the case of the hypotheses full-siblings or avuncular. Statistical evaluation was computed through the quantification of likelihood ratios, assuming the genotypic configuration of 100,000 simulated families, for each of the kinship problems analyzed: full-siblings vs. unrelated, full-siblings vs. half-siblings, half-siblings vs. unrelated, avuncular vs. unrelated, first cousins vs. unrelated, and half-first cousins vs. unrelated. For each of the six studied cases we have obtained results allowing us to weight the informative power impact of increasing the number of markers and of the addition of an extra individual, considering (a.) the use of different sets of STRs (from 8 to 35), and (b.) the introduction of a third undoubted relative. Based on these results we were able to provide recommendations for each case both on the minimal number of STRs to be used and on the third relative whose genetic analysis should be privileged. [ABSTRACT FROM AUTHOR]