Read length distribution of whole genome sequnces of human

Question

[Yellapu]

Hello,

I am working on deriving HRD scores for human geome samples.

As part of this, the quality report after adaptor triming showsa broad range in the reads length distribution. I am attaching the image here.

Do I require to filter the reads of specific length before proceeding for alignmnet with reference (or) is it good to proceed for alignmnet as is? I could see there is an intense length distribution below 150 bp and majority of the reads are around 150 bp. What I suspect is to filter the reads below 150 bp?

Please anyone give your thoughts.

Your inputs are greatly appreciated.

[Ram]

This is some urgent help.

This kind of statement does not help. On a forum powered by volunteers' time, pushing your sense of urgency has an effect of throwing people off as people might move on to something they can actually help with instead of dealing with more pressure during their time off their actual jobs.

Answer 1

[GenoMax]

It is normal to see a range of read lengths after adapter trimming, especially if you have short inserts in your libraries. You could try filtering the reads to be of a certain minimum length (we started next generation sequencing data with ~35 bp reads and those can still be reasonably aligned to human genome). Very short reads will tend to multi-map. You will need to decide what to do with those. Looks like you have millions of reads here so filtering out reads shorter than 50 bp may end up losing a relatively small percentage of data.

[Yellapu]

Thanks GenoMax for providing a reasonable answer. I am in the process of checking the alignment rate before and after filtering. I prefer not losing millions of reads actaully and so including the reads >35 bp. Thanks