Different alignments rates with bwa mem (0%) and bowtie2 (82%)

Question

[James Ashmore]

I have ChIP-seq data which has been trimmed using trim galore (all default settings). When I align these trimmed reads using bwa mem (all default settings) I get no aligned reads, but when I use bowtie2 (all default settings) I get an 82% alignment rate. The trimmed reads are ~20 bp in length. The run time with bwa is a lot faster than bowtie2 so I think maybe bwa is discarding reads based on some internal quality or length filter.

Below is the samtools flagstat output for bwa mem:

4476553 + 0 in total (QC-passed reads + QC-failed reads)
0 + 0 secondary
0 + 0 supplementary
0 + 0 duplicates
0 + 0 mapped (0.00% : N/A)
0 + 0 paired in sequencing
0 + 0 read1
0 + 0 read2
0 + 0 properly paired (N/A : N/A)
0 + 0 with itself and mate mapped
0 + 0 singletons (N/A : N/A)
0 + 0 with mate mapped to a different chr
0 + 0 with mate mapped to a different chr (mapQ>=5)

and for bowtie2:

4476553 + 0 in total (QC-passed reads + QC-failed reads)
0 + 0 secondary
0 + 0 supplementary
0 + 0 duplicates
3696458 + 0 mapped (82.57% : N/A)
0 + 0 paired in sequencing
0 + 0 read1
0 + 0 read2
0 + 0 properly paired (N/A : N/A)
0 + 0 with itself and mate mapped
0 + 0 singletons (N/A : N/A)
0 + 0 with mate mapped to a different chr
0 + 0 with mate mapped to a different chr (mapQ>=5)

[vmicrobio]

you may find some interesting suggestions here When and why is bwa aln better then bwa mem? and here http://seqanswers.com/forums/showthread.php?t=15200

[stolarek.ir]

try mapping only the longest reads. BWA mem has default minimum seed length as 19. But anyways mem is not the right option for very short reads, I would stick with the bwa aln if not bowtie2

[5heikki]

bwa mem default minimum seed length is 20 bp (or 19 bp?). If none of your reads exceed that length, then the seed alignments obviously can't be extended..

[James Ashmore]

According to the default, the seed length is 19bp. The minimum read length I have is 20bp and the longest is 26bp. I'll try a seed length of 10 and see if that increases the alignment rate.

Edit:

No improvement setting the initial seed length to 10 bp. Still a 0% alignment rate with bwa.