Read showing Insertion of 41 bases when aligned with BWA

Question

[kspata]

Hi All,

I aligned reads of a viral vector using both Bowtie2 local and BWA-MEM. 1300 reads are mapped to this position.

I observed that BWA shows an insert of 41 bases (in the ITR region), the mapping quality of this read is 60, AS:i:91 and CIGAR string of 53S91M41I7M27I4M.

While bowtie2 local does not show this variant and Mapping Quality is 60, AS:i:182 and CIGAR string 53S91M79S.

I believe bowtie2 softclipped these bases while BWA introduced insert to obtain higher alignment score.

I dont believe this variant is true. Can you please help me understand this, which alignment should I consider to make accurate variant call?

Thanks in advance !!

Answer 1

[swbarnes2]

So you have a 223 bp long read, and you think only 91 bases of it is real?

More clipping is a simpler answer than what bwa did, but why do you have so much cruft on the ends of your reads to be trimmed?

[kspata]

Hi swbarnes2,

I dont understand what you mean to say. I understand BWA is introducing inserts instead of soft clipping. I dont know why this is happening. Also, this region is ITR and contains CCCGGG repeats.

Does BWA forcefully introduce inserts to obtain high alignment score?

[GenoMax]

How many reads do you have that are spanning this region (i.e. depth of support for this observation)?

[swbarnes2]

Why does your read need so much end clipping? Do all your reads need that much clipping?

[kspata]

The total depth at this location is 1300 (Reference Reads + Alternate Reads). The variant is found at 10% frequency. This reads map to the ITR regions. Some bases are automatically softclipped by the aligners. Also, the variants are located in the reverse reads.