Bowtie 2 alignment

Question

[Elizabeth]

Hello All, My objective is to align paired reads against my reference genome (nematode) and I have used bowtie2 and bwa to accomplish this. The process is complete, but the alignment stats from bowtie2 doesn't make any sense.

Initially I used the --very-sensitive parameter, hence I figured removing this would generate a better alignment. But, the results are the same.

Below is a snapshot of the stats. Now, I extracted the top 10 unmapped reads and blasted it on NCBI. 9/10 are positive hits for nematode rDNA regions. I'm assuming the person who did the assembly excluded the rDNA portions. But, what I do not understand is this portion should be relatively smaller. 33% is a good chunk of the genome which cannot be just the rDNA cluster. I did trim the reads, remove the adapters, remove the contaminants (bacteria,archaea,viral, human) using kraken before doing the alignment. I'm not sure what's going wrong here. Please help. Thank you.

143858045 reads; of these:
143858045 (100.00%) were paired; of these:
      47485448 (33.01%) aligned concordantly 0 times ## this right here
      68223621 (47.42%) aligned concordantly exactly 1 time
      28148976 (19.57%) aligned concordantly >1 times

Answer 1

[Istvan Albert]

If I recall correctly to make bwa and bowtie2 comparable you have to use the --very-sensitive-local parameter with bowtie2

it has to do with the ignoring the soft clipping penalty - so try this new parameter first and see if that changes anything

[Elizabeth]

Thank you so much. I'm getting comparable results now. Appreciate the quick response.