Evaluation of haplogroup prediction softwares

CAPUTO M; BOILLO C; ALECHINE E; SALA A; CORACH D

Innsbruck

Congreso; DNA in Forensics 2012. "DNA in Forensics: Exploring the Phylogenies"; 2012

Y User and EMPOP

In population genetics studies, huge collections of Y-STR haplotypes have been published; unfortunately most of them lack information concerning the haplogroups to which the haplotype belong. Inferring haplogroups from haplotype data is an attractive alternative, however a robust informatic approach should be used. Based on the most theoretical criticism rised by some authors, we decided to empirically test the Athey’s Haplogroup Predictor with samples previously typed for haplogroup defining SNPs. One hundred and one unrelated male samples were haplotyped by AmpFLSTR Y-Filer (Applied Biosystem) and haplogrouped by SNP detection performed by different approaches, namely Real Time PCR followed by High Resolution Melting, amplicon sequencing, SnapShot mini-sequencing and/or primer specific PCR. The analyzed SNPs were M3 for Q1a3a, M269 for R1b1b2 and U179 for I Haplogroups. Thirty one samples were assigned as “not determined” and 70 samples were haplogrouped as I, Q1a3a or R1b1b2. The prediction of Y-Chromosome haplogroups from these samples was performed by Athey’s Haplogroup Predictor considering “equal priors”. All the haplogroups predicted by the online program matched the assignment by SNP genotyping. An average probability of 99.7%, 96.9% and 99.99% was obtained for Q, I, R1b haplogroup assignment, respectively, using Y-Filer STR panel.    An adequate correlation between the haplogroups predicted by the software and SNP haplogroup typing was obtained. The use of Athey’s Haplogroup Predictor by means of 17 markers (Y-Filer Panel) showed a reliable accuracy to predict at least I, Q1a3a and R1b haplogroups. Its use might offer an opportunity for retrieving valuable information from published haplotypes. Increasing the number of reference samples from which haplogroup have been precisely defined by SNP analysis will highly improve the software accuracy. Meanwhile, this approach represents an acceptable screening criteria  that may allow analyzing previously published results.