Genetic mapping of 15 human X chromosomal forensic short tandem repeat (STR) loci by means of multi-core parallelization

Toni Marie Diegoli, Heinrich Rohde, Stefan Borowski, Michael Krawczak, Michael D. Coble, Michael Nothnagel

Research output: Contribution to journalArticlepeer-review

15 Scopus citations


Typing of X chromosomal short tandem repeat (X STR) markers has become a standard element of human forensic genetic analysis. Joint consideration of many X STR markers at a time increases their discriminatory power but, owing to physical linkage, requires inter-marker recombination rates to be accurately known. We estimated the recombination rates between 15 well established X STR markers using genotype data from 158 families (1041 individuals) and following a previously proposed likelihood-based approach that allows for single-step mutations. To meet the computational requirements of this family-based type of analysis, we modified a previous implementation so as to allow multi-core parallelization on a high-performance computing system. While we obtained recombination rate estimates larger than zero for all but one pair of adjacent markers within the four previously proposed linkage groups, none of the three X STR pairs defining the junctions of these groups yielded a recombination rate estimate of 0.50. Corroborating previous studies, our results therefore argue against a simple model of independent X chromosomal linkage groups. Moreover, the refined recombination fraction estimates obtained in our study will facilitate the appropriate joint consideration of all 15 investigated markers in forensic analysis.

Original languageEnglish
Pages (from-to)39-44
Number of pages6
JournalForensic Science International: Genetics
StatePublished - 1 Nov 2016


  • Forensic analysis
  • Genetic linkage
  • Genetic mapping
  • High performance computing
  • Short tandem repeat
  • X chromosome


Dive into the research topics of 'Genetic mapping of 15 human X chromosomal forensic short tandem repeat (STR) loci by means of multi-core parallelization'. Together they form a unique fingerprint.

Cite this