Dense DNA Data for Enhanced Missing Persons Identification

DNA profiles from human remains are compared with potential family reference samples (biological relatives such as parents, offspring, siblings, cousins) and indirectly identified by kinship analysis. Indirect comparisons are based on leveraging identity by descent (IBD) between related individuals. In many situations, not all immediate family members are available; at times only a single reference sample is available and it may derive from a single sibling, a grandparent, an uncle or aunt, a half sibling or a first cousin. Presented with only genotypes of these more distant reference individuals it often is difficult to develop statistically significant supported associations. Moreover, some related individuals can appear to be unrelated (or uninformative), thus reducing the weight substantially and potentially supporting an exclusion as well as some unrelated individuals may falsely appear related. There is uncertainty about the findings because of the stochastic nature of allele sharing between relatives. This situation is exacerbated for second degree relatives and beyond. Typically, up to third degree relatives are sought as reference samples by investigators for missing persons identifications.

Genetic genealogy tools employ dense DNA testing (SNP microarrays chips) to determine biological relationships between individuals which may be used to help investigations. If a weak or uninformative association is made due to limited STR data, a SNP panel (using a SNP chip or by NGS) could be typed to further support or refute the limited association. As the marker densities are generally high, the precision is similarly high, allowing for tight confidence interval estimation on related samples. This approach should be invaluable for missing persons identifications as correct kinship inferences can be made, and equally important exclusions can be inferred. The proposal herein will develop strategies to extract reliable and accurate data from dense SNP typing results from less than optimum input DNA, i.e., the type of samples encountered in human remains. Through simulation and empirical studies, the feasibility of inferring kinship associations up to 3 degrees distant will be determined. If successful (preliminary data already support feasibility), current requirements for family reference samples can be relaxed reducing the need for a maternal (or paternal relative) and a first degree relative reference sample. A single reference sample (as distant as first cousins and likely beyond) would be sufficient for effecting associations as well as reduce the rate of false and negative associations that occurs with current DNA typing systems. Moreover, the identification success rate will increase.

Note: This project contains a research and/or development component, as defined in applicable law, and complies with Part 200 Uniform Requirements - 2 CFR 200.210(a)(14).

CA/NCF