An iterative algorithm for tagSNP selection based on information entropy analysis

In this research, an iterative algorithm based on information entropy analysis is proposed for tagSNPs selection. Dynamic programming algorithm is employed to partition haplotypes into blocks with the first constraint, the minimum total block entropy. Missing SNPs are inferred with the second constraint, the minimum number of tagSNPs. The proposed algorithm iterates between these two phases with the above two constraints until the number of tagSNPs reaches its minimum. The proposed algorithm is simulated with two data sets, including Daly et al. 2001, and Patil et al. 2001. Experimental results show that the proposed scheme reduces the number of tagSNPs required in haplotypes significantly.