Comparative analysis of algorithms for LD tagSNPs selection in a single population

To reduce genotyping costs, tagSNPs selection has gained lots of research interest and several algorithms have been proposed for LD tagSNPs selection. However, there is little understanding of benefits and drawbacks underlying the proposed approaches. In this paper, we broadly analyzed the existing algorithms for LD tagSNPs selection in a single population. It can be concluded that the original greedy algorithm is easy to implement and TAGster refines the selection steps of the original greedy algorithm. Results suggested that HTag can receive the highest tagging efficiency. In addition, we found that the precinct partitioning strategy can dramatically reduce the runtime of TAGster and HTag. Theoretical analysis revealed that unstable numbers of tagSNPs may be received by these methods sometimes.