Aligning read with non-genome encoded sequence

Question

[biplab]

I have pair-end RNA seq data containing variable length of non-genome encoded sequence, what will be the best way to align them to reference genome? Thank you so much.

[GenoMax]

non-genome encoded sequence

Can you clarify what that means? Is this an artificial insert that was introduced into the genome?

[biplab]

Sequencing reads came from 3' end of mRNA that contain polyA tail in R2 file. However R1 does not have any non-genome encoded sequence and align properly to the genome. Here non-genome encoded sequence is the polyT stretches in R2 reads.

[GenoMax]

Those should be easy to trim using bbduk.sh with literal=AAAAA option. bbmap.sh will also be able to softclip those tails as long as there is enough sequence present to allow a reasonable match. Both programs are part of BBMap suite.

[biplab]

Thank you so much. I will look into BBMap.

Answer 1

[Devon Ryan]

Either trim off the polyA tail on read 2 or use STAR (or another local aligner) and allow a lot of soft-clipping. I suspect that trimming read 2 a (then aligning the remaining pairs and orphans) will produce the best results.

[biplab]

Thanks you so much. I will try soft-clipping and trimming to see which one work best. There might be some non-A bases in polyA tail because of sequencing error, what percent of error should I consider during trimming polyA tail?

[GenoMax]

With bbduk.sh you can remove all sequence to the right of the poly-A tail once it is located. So it does not matter what sequence may be there.