I am about to commit some serious time to analyzing data produced on a SOLiD 5500 and I had a few questions about sequence quality. First off, does quality matter that much if I have >10 million reads per sample? Second, does quality seriously alter mapping with TOPHAT (RNAseq analysis)? I would think this is less of a problem then if I was doing SNP detection or something. Finally, I can filter the reads with quality scores below 10 easily in GALAXY, but then I lose ~55% of the reads! Seems like a lot. Here are some images pre and post filtering. Am I just trying too hard to clean up the data or should I filter? Thanks in advance.
