already genotyped marker to act as a proxy. If a SNP does not have any markers in high LD with it, it cannot accurately be predicted. This means in practice that well-covered regions can easily be filled in but poorly covered regions cannot be improved to any major extent. Given that imputation methods utilise the haplotype structure of reference datasets for the inference of genotypes, The 1000 Genomes Project will also contribute highly to our understanding of genotype imputation methods and their accuracy for imputation of polymorphisms of different frequencies.
