Disable iterative search in BWA aln

[yanasmann]

Hi Heng Li, I tried to align 50-bp Illumina Short Reads to RefSeq RNA. Because I expected that some reads will be mapped to multiple transcripts from the same gene, I used the -N parameter to gather all possible mappings. My command is as following:

bwa aln -N -l 32 -t 4 Indexed_RefSeq FASTQ > xxx.sai

However, when I exam the output sam files, I found the following example, where the MAPQ=0, and CIGAR = 50M. My experience is this means that the read mapped to multiple locations. So the -N was not activiated? Please HELP? Thank you!!!

HWI-EAS217:1:1:222:1093#0 16 gi|57863262|ref|NM_198393.2| 4097 0 50M * 0 0 CCATCTCAGGAGCTACTTGATGACATTGAGCTCTTGAAACAGCAGCAGGG [_N]Nbab_aaa[aa]babaaabaababaabaaabaabaaabaaa XT:A:R NM:i:0 X0:i :2 X1:i:0 XM:i:0 XO:i:0 XG:i:0 MD:Z:50

[yanasmann]

We digged a little more into this. It seems that BWA -aln -N gives all alignment locations. However, the samese doesn't write out all the location that mapped equaly well.

Is that correct? Heng? Do we have options to read the .sai files directly to extract that information?