adarob / eXpress

Streaming fragment assignment and quantification for high-throughput sequencing.
bio.math.berkeley.edu/eXpress
Other
36 stars 10 forks source link

softclipped parts of reads as considered as indels #11

Open pavanvidem opened 8 years ago

pavanvidem commented 8 years ago

eXpress seems to treat soft clipping in read as indels. This can be proven with the option --max-indel-size. I have reads with soft clips ranging from 3 to 40 bases. With default --max-indel-size (i.e., 10), express could only quantify reads with max indels <= 10 includin soft clips. I lost most of my perfectly matching reads in quantification because they have soft clipped parts which are longer than 10 bases long . With changing the --max-indel-size to 40 could quantify most of the reads, but sometimes lead to segfault.