This is a test version of Biostars. For the public version, visit https://www.biostars.org.
behaviour of bowtie compared to junction aligners?

Hi,

I have human RNASeq data generated using lexogen kit. I have followed their analysis protocol (to process raw data). Then used tophat2 for alignment, observed only 53%. check unalinged using blast they are going human rRNA sequences end2end. Then i used hisat2, the % increased to 75. Then tried bowtie2 end2end parameter (which does not align junction read i assume) has given 79%.

Unaligned reads from different aligners checked against NCBI NT using BLAST they are hitting to human rRNAs/ncRNAs. Questions,

  1. How bowtie is preforming better than junction aligners?

  2. How do i improve the overall alignment?

Reference used was GRCh38 from ensemble

I have seen the same trend for fungi sample as well.

Any in site what could be the reason?

rna-seq alignment
• 2,854 views • link
updated by Ram Ram • written by popayekid55 popayekid55

Did you use equivalent settings for each aligner? Could you post the command-line for each aligner?

• 0 views • link
written by h.mon h.mon

used default parameters, bowtie2 end-end, same in tophat enabled b2-sensitive.

• 0 views • link
written by popayekid55 popayekid55

1 answer

  1. It shouldn't, since it'll miss everything that's spliced
  2. Use STAR or another mapper that allows local alignment. You should get 90+% alignment.
• 0 views • link
written by Devon Ryan Devon Ryan

I think you missed this part.

Unaligned reads from different aligners checked against NCBI NT using BLAST they are hitting to human rRNAs/ncRNAs.

popayekid55 : Not much you can do if above is true.

• 0 views • link
written by GenoMax GenoMax

I saw that, my presumption is that the edit distance is high enough that the reads aren't aligned by the aligners used.

• 0 views • link
written by Devon Ryan Devon Ryan

Or the reference did not contain rDNA repeat.

• 0 views • link
written by GenoMax GenoMax

That's usually part of the human reference, though if this were mouse I'd agree.

• 0 views • link
written by Devon Ryan Devon Ryan

Unaligned reads from different aligners checked against NCBI NT using BLAST they are hitting to human rRNAs/ncRNAs.

Reason I thought that the reference here must not have rDNA.

• 0 views • link
written by GenoMax GenoMax

Not much you can do if above is true.

how bowtie is performing better than splice aligners?

• 0 views • link
written by popayekid55 popayekid55

TopHat2 generally uses bowtie2. Are you sure you were using bowtie v.1?

• 0 views • link
written by GenoMax GenoMax

i have used bowtie2-2.2.5 version. Same version was imported to tophat as well

• 0 views • link
written by popayekid55 popayekid55

Log in to answer this question.

More posts like this