Hi all,

I am using BWA mem on samples that have been sequencend on an Illumina 
HiSeq2500 in paired-end 2*125 nuclotides.
I have not change any default option.So, in the results, I was expecting not to have any read with AS tag lower than 30, as this is the default value for -T option. The fact is that I have many reads with AS lower than 30 (most of them are 19, 20 or 21). These low alignment scores coincide with reads having a lot of soft-clipped sequences. Below is an example of read like that in my results:

HWI-D00562:34:C6KW3ANXX:5:2205:14887:84822    99    chr12 71309226    
40    17S19M89S    =    7130922619 
CCAATGAACCATATCAGTATTCCTGTGTCTCTTAAGCCACCTTATCACCCCTGTTGATATAGCAATGTGCTGTCTCTTATACACATCTCCGAGCCCACGAGACTAAGGCGAATCTCGTATGCCGT 
CCCCCGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGFGGGGGGGGGGGEGGGGGGGGGGGGGGGGG<CEEGGGGG<EGGGEGGGGGG..<FGGGGGGGGG 
MD:Z:19    PG:Z:MarkDuplicates    NM:i:0    MQ:i:40    AS:i:19 XS:i:0
HWI-D00562:34:C6KW3ANXX:5:2205:14887:84822    147    chr12 71309226    
40    73S19M33S    =    71309226-19 
CACCGAGATCTACACTAGATCGCTCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCAATGAACCATATCAGTATTCCTGTGTCTCTTAAGCCACCTTATCACCCCTGTTGATATAGCAATGTG 
GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGEFGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGBCCCC 
MD:Z:19    PG:Z:MarkDuplicates    NM:i:0    MQ:i:40    AS:i:19 XS:i:0

Why is this read pair reported in the output of BWA mem? It should have been 
discarded thanks to the -T 30 option, isn't it? How can I fix that?

Thank you!