Hello,

I'm trying to align 75bp reads from a .fastq file (~1G, sequencing of MtDNA only), but due the library preparation design (not going to cover details), bwa mem is not able to align most of the reads. The problem with my reads is that, for every read, only ~45bps will be matched to reference, all other bases are junk and have no match to MtDNA reference.

It seems like if within this 45 bps, 2 or more mismatches occurred NOT continuously, bwa mem will simply think there is no good match worth recording. For example, one read looks like:

@NS500323:104:H3L2KBGXY:1:11101:8802:1033 1:N:0:TAAGGC
TGGTTNTGTTGATTCAAATTATGTGTTTTTTGGAAAGTCATGTCAGATCGGAAGAGCACACGTCTGAACTCCAGTC
+
A//AA#6EEE6EEEEEEEEA6EEEEEEEEEEEEEEE6EEEEAEEEEEEEEEEEAE/EAEEEEEEE/A<EEE</EEE

A simple blat alignment can give me (there is on "n" to "g" and one "t" to "c" mismatches):

00001 tggttntgttgattcaaattatgtgttttttggaaagtcatgtcag 00046
<<<<< ||||| ||||||||||||||||||| |||||||||||||||||||| <<<<<
10845 tggttgtgttgattcaaattatgtgctttttggaaagtcatgtcag 10800

but bwa mem will not have it in the sam file.

I'm wondering if I should consider switch to blat standalone version, or if bwa-backtrack. Any suggestion please? Thank you! I suspect it was due to the default setting of mismatch penalty, so I try to change the "-B" option of bwa, but it didn't help.