Accurate imputation of rare and common variants in a founder population from a small number of sequenced individuals

Genet Epidemiol. 2012 May;36(4):312-9. doi: 10.1002/gepi.21623. Epub 2012 Mar 28.


Advances in DNA sequencing technologies have greatly facilitated the discovery of rare genetic variants in the human genome, many of which may contribute to common disease risk. However, evaluating their individual or even collective effects on disease risk requires very large sample sizes, which involves study designs that are often prohibitively expensive. We present an alternative approach for determining genotypes in large numbers of individuals for all variants discovered in the sequence of relatively few individuals. Specifically, we developed a new imputation algorithm that utilizes whole-exome sequencing data from 25 members of the South Dakota Hutterite population, and genome-wide single nucleotide polymorphism (SNP) genotypes from >1,400 individuals from the same founder population. The algorithm relies on identity-by-descent sharing of phased haplotypes, a different strategy than the linkage disequilibrium methods found in most imputation algorithms. We imputed genotypes discovered in the sequence data to on average ∼77% of chromosomes among the 1,400 individuals. Median R(2) between imputed and directly genotyped data was >0.99. As expected, many variants that are vanishingly rare in European populations have risen to larger frequencies in the founder population and would be amenable to single-SNP analyses.

Publication types

  • Research Support, N.I.H., Extramural

MeSH terms

  • Algorithms
  • Computational Biology / methods*
  • Ethnicity
  • Exome
  • Genetic Variation
  • Genome-Wide Association Study
  • Genotype
  • Haplotypes
  • Humans
  • Linkage Disequilibrium
  • Pedigree
  • Polymorphism, Single Nucleotide
  • Religion
  • Reproducibility of Results
  • Risk Factors
  • Sequence Analysis, DNA
  • South Dakota
  • White People / genetics