Yale geneticist urges committee to adopt microhaplotypes as next-generation forensic markers

Dr. Ken Kidd, professor emeritus of genetics at Yale School of Medicine, presented research arguing that microhaplotypes — compact DNA segments containing multiple single-nucleotide polymorphisms — offer superior performance over conventional short tandem repeat (STR) markers for several forensic tasks, including mixture deconvolution and individualization. The presentation concluded with a recommendation that a committee adopt a documented set of microhaplotypes as an initial standard for an offender database. The presentation date and venue were not specified in the transcript.

Kidd told attendees that microhaplotypes can support five forensic uses in one marker set: individualization, ancestry inference, phenotype inference, mixture deconvolution and familial relationship inference. He defined a microhaplotype as "just a small segment of DNA with 3 or 5 or 6 SNPs within it, but small enough that with massively parallel sequencing, one can get a sequence read across the whole interval" and described how phased sequencing reads produce multiallelic haplotypes with minimal stutter and very low mutation rates compared with STRs.

The research team has identified 182 candidate microhaplotype loci and characterized them on 83 populations, with more than 5,000 fully characterized individuals in the current analysis, Kidd said. He described ranking roughly 132 candidate loci across those populations by two statistics — effective number of alleles and informativeness — to select markers for different forensic purposes, noting that the top 50 loci show an average effective number of alleles around 3–3.5 and informativeness values up to about 0.3 in their metrics.

Kidd presented random-match probability calculations for several panels, including a previously published 55-SNP ancestry panel and a 45-SNP identity snip panel. He said the 45 identity snips produced probabilities on the order of 10^-15 in some population calculations, and that other selected loci produced values as small as 10^-30 to 10^-47 when plotted on a negative log10 scale, but he cautioned that those figures depend on correction factors and assumptions and are not the final word on statistical certainty.

Laboratory evaluations are under way with industry and academic partners, Kidd said, naming Thermo Fisher and Daniel Podini's lab at George Washington University. He gave a mixture example from a 36‑plex run in which a particular four‑SNP haplotype was observed 192 times in the reads, indicating a four‑person mixture that could be disentangled from the sequence data — a result he contrasted with STR profiles that in the same amplifications were often uninterpretable.

Kidd criticized limitations of STRs and of current probabilistic genotyping approaches, saying they suffer from stutter and confounding signals that make interpretation difficult in compromised samples. "So, I strongly believe that microhaplotypes are the next generation of forensic markers," he said, and urged the committee to decide on a well-documented initial set that could be used to design an offender database. He also acknowledged ongoing needs: reference databases of offender profiles and continued evaluation of marker independence and linkage.

Kidd thanked collaborators and "the thousands of individuals who have allowed us to study their DNA," and acknowledged funding from the National Institute of Justice for molecular work and database development over nearly a decade. After his remarks the moderator invited questions from the audience.