Rxivist logo

The rate of false polymorphisms introduced when imputing genotypes from global imputation panels

By Ida Surakka, Antti-Pekka Sarin, Sanni E Ruotsalainen, Richard Durbin, Veikko Salomaa, Mark J. Daly, Aarno Palotie, Samuli Ripatti, SISu project group

Posted 13 Oct 2016
bioRxiv DOI: 10.1101/080770

Previous studies1,2 have shown that large multi-population imputation reference panels increases the number of well-imputed variants. However, to our knowledge, no previous studies have evaluated the rate of introduced variation in monomorphic sites of the study population when using imputation panels with admixed populations. In this study we evaluate the rate of false positive variants introduced by the imputation of Finnish genotype data using global reference panels (Haplotype Reference Consortium1; HRC, and the 1000Genomes project Phase I3; 1000G) and compare the results to a Finnish population-specific reference panel combining whole genome and exome sequenced samples. In sites that were monomorphic in our test set, we observed high false positive rates for the global reference panels (4.0% for 1000G and 2.6% for HRC) compared to the Finnish panel (0.26%). This rate was even higher (7.4%) when using a combination panel of 1000G and Finnish whole genome sequences with cross-panel imputation.

Download data

  • Downloaded 983 times
  • Download rankings, all-time:
    • Site-wide: 24,669
    • In genetics: 1,134
  • Year to date:
    • Site-wide: 109,815
  • Since beginning of last month:
    • Site-wide: 110,692

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide