Haloplex HS - should you remove duplicates?

Question

[amy__]

Hello,

I am aware a question has been asked previously but quite a few years ago.

I have haloplex hs data, I have done:

1. trimming with Agilent surecall trimmer
2. cutadapt to remove any remaining adaptors present
3. bwa mem
4. samtools view
5. samtools sort

However, in the Agilent surecall workflow they recommend using their tool locatit to remove duplicates.

When I do this:

1. trimming with Agilent surecall trimmer
2. cutadapt to remove any remaining adaptors present
3. bwa mem
4. samtools view
5. locatit
6. samtools sort

I am left with a significantly reduced number of reads for calling. I am aware it is probably the duplicates that are the problem here, but I have also seen other papers where people do not remove the duplicates? Does anyone have any suggestions?

I also need to merge my bams so that might increase the reads slightly but not by much.

Thanks! Amy