Fig. 2
From: An effective method to resolve ambiguous bisulfite-treated reads
Comparison with random selection method on the real human data set. The data set we choose is SRR901388. After using Bismark to align all BS-reads to the reference genome, all possible alignment positions of multireads are obtained. We divided multireads into several groups according to the numbers of alignment positions to explore the effect of different methods on each dataset. The horizontal ordinate is the number of multiread’s mapping positions. The ordinate in the figure means the proportion of multireads correctly aligned by different methods