NGS data with >50% duplicates

[dingxiaofan1]

Does anybody know how to deal with data like this. What's worse, all of the data was generated around repeat elements. Alignment was good, but rmdup could reduce >50% data, and snp calling has a very high false positive rate.

[adaptivegenome]

New library prep?

[dingxiaofan1]

New library prep?

Thank you, but I donot quite understand. This is the data they asked me to analyze, so I cannot improve the result by modifing experiment conditions.

[harryzs]

you should first tell us what kind of data you have??

NGS data for ChiP-seq, RNA-seq?

or

see this http://seqanswers.com/forums/showthr...cates+chip-seq

[kopi-o]

50% duplicates is not at all uncommon, it is impossible to tell whether it is worrying or not without details about how the data were generated.

[dingxiaofan1]

you should first tell us what kind of data you have??

NGS data for ChiP-seq, RNA-seq?

or

see this http://seqanswers.com/forums/showthr...cates+chip-seq

Something like target specific sequence PCR sequencing, but the target is not unique.

[dingxiaofan1]

50% duplicates is not at all uncommon, it is impossible to tell whether it is worrying or not without details about how the data were generated.

Yes, the phenomena is not uncommon. But in my case duplicates could leads to very serious misalignment problems and genotype calling errs.

[kopi-o]

Then I would suggest to remove the duplicates with Picard MarkDuplicates.