High duplication percentage

[klebsiella]

Dear all,

I know that this thread was already asked before, however I wonder if someone has a new/better explanation/suggestions.

I am encountering a high duplication level in my chip-seq libraries. almost 99% of the reads are eliminated when considering only one read to map with Bowtie.

I am starting my libraries with around 2/3 ng of IP materials. I am using the NEB kit for the library prep. The library concentration is not that high too.
I am doing 18 cycles. I am multiplexing 10 samples at a time.
Inputs sequencing seems better with around 40% of duplication.

Any suggestion is the most welcome,

Best,

[huguesparri]

You can try to size select after the PCR enrichment rather than before. It tends to increase the diversity in the final libraries.

[klebsiella]

Dear Huguesparri,

Many thanks for your reply,
Indeed, I will try now either 1) avoiding completely the first size selection and put everything into the PCR or 2) making a really big size selection first (i.e 200-800 bp) and go with that to the PCR.
I think that I will go for the second option. I believe that in both cases this will help to increase the PCR complexity and get better results (less duplication).

PS: Also I will also try to start with more materials at the beginning,

Best