Fig. 2

The performance of different methods in removing mapping bias from simulated reads. A Performance of Hisat2 or BWA, either alone (e.g., Hisat2-only or BWA-only as indicated) or in combination with WASP (e.g., BWA_WASP) on the simulated reads with different lengths (e.g., 36 bp, 50 bp, and 100 bp, as indicated). Two types of datasets were tested: either single-end (SE) reads with different lengths: SE-36 bp, SE-50 bp, and SE-100 bp; or paired-end (PE) reads with different lengths: PE-50 bp and PE-100 bp. B and C. Performance of the indicated methods, as detailed in (A), on simulated single-end (B) and paired-end (C) reads with a fixed read length of 100 bp and difference effect sizes. ‘Hisat2_only’ and ‘BWA_only’ referred to methods with basic parameters without removing mapping bias, while ‘Hisat2_SNP’ adopted the parameter of ‘-snp’. Using BWA to map and remap and using WASP to remove bias were indicated with "BWA_WASP". The mean square error (MSE) was used to measure the performance of the different methods. Effect size refers to the ratios of the reference allele and alternative allele with 10% of SNPs between 1.5:1 and 5.0:1