although similar issue been reported several times but i tried all the suggestions but problem still exist. Im working with RRBS data obtained from Cow sperm DNA. Standard illumine adapters and low quality sequences were removed from pair reads using –paired option in trim galore. As I can see in fastQC report, read 1 is ok but read 2 is not good as read 1 and for all samples read 2 has a lower quality compared to read 1. In Bismark and using Bowtie 2 alignment, I tried pair end alignment first (PE) with different --Score-min adjustments and surprisingly the mapping efficiency was very low. Following some suggestions in different forums, I tried with –pbat option and furthermore I tried to align the reads separately, as well. But as you can see bellow, there is a huge difference in mapping efficiency between read 1 and 2 and generally pairwise alignment mapping efficiency is very low. In addition, I realized that if in single end alignment I include –pbat, for read 1 and 2 mapping efficiency decreased and increased, respectively. Do you think that my read 2 is causing the problem in pair end alignment? Any suggestion for improving the mapping efficiency greatly appreciated.
--Score-min Default (L, 0, -0.2)
PE: 11.4 % PE with --pbat: 0 % SE, R1: 27.3 % SE, R2: 0 %
--Score-min (L, 0, -0.6)
PE: 19.2 % PE with --pbat: 0 % SE, R1: 29.3 % SE, R2: 0.1 %
--Score-min (L, -0.6, -0.6)
PE: 19.2 % PE with --pbat: 0 % SE, R1: 29.4 % SE, R2: 0.1 %
--Score-min (L, -0.6, -1)
PE: 29.5 %
PE with --pbat: 0.3 %
SE, R1: 30.7 % SE, R1 with --pbat: 16.4 %
SE, R2: 0.7 % SE, R2 with --pbat: 28.7 %
--Score-min Default (L, 0, -0.2)
PE: 11.4 % PE with --pbat: 0 % SE, R1: 27.3 % SE, R2: 0 %
--Score-min (L, 0, -0.6)
PE: 19.2 % PE with --pbat: 0 % SE, R1: 29.3 % SE, R2: 0.1 %
--Score-min (L, -0.6, -0.6)
PE: 19.2 % PE with --pbat: 0 % SE, R1: 29.4 % SE, R2: 0.1 %
--Score-min (L, -0.6, -1)
PE: 29.5 %
PE with --pbat: 0.3 %
SE, R1: 30.7 % SE, R1 with --pbat: 16.4 %
SE, R2: 0.7 % SE, R2 with --pbat: 28.7 %
Comment