Tools to predicting binary states on the human Y chromosome from STR data

A novel Bayesian algorithm, “WYZARD”, is designed to predicting binary states on the human Y chromosome from STR data. It allows users to retrieve linkage probabilities between combination of alleles at the 8 most widely used STR loci (DYS19, DYS389I, DYS389II, DYS390, DYS391, DYS392, DYS393, DYS388) and the derived mutations defining 1 super-haplogroup [F(xK)], 4 haplogroups (I, L, N, Q) and 14 sub-haplogroups (E3a, E3b1a, E3b1b, E3b3, G1, G2, I1a, I1b, I1c, J1, J2, R1a, R1b, R2), which encompass 99% of West Eurasian variability. Prior probabilities were calculated from a geographically unbiased repository of 3,672 chromosomes we collected from published and unpublished sources. The robustness of the WYZARD approach and of other six approaches of haplogroup assignment following distance-, Bayesian- and frequencybased methods was assess by comparing predictions against the true haplogroup of 135 haplotypes with Austrian origin. Incorrect assignments ranged between 11.1% and 16.3%, with WYZARD giving the lower values among Bayesian methods (14.8 %). Being misleading results limited to few couples of haplogroups, a 100% rate of correct assignments can be reached introducing STR-based predictions in routine protocols for Y binary screening. It would shortcut the diagnosis of binary mutations with costs 50-70% lower than standing-alone approaches.