An overview of phasing methods for polyploids

We simulated unphased genotype data with various numbers of individuals and SNPs (10 repetitions). We evaluated the performance of both the methods, using switch error rate as shown in the plot below. Switch error per individual is the minimum number of switches made to reconstruct its true haplotypes. On smaller data sets, SATlotyper outperformed Polyhap, but it failed to converge even after three days for larger ones.