Poly AT-stretches in ChIP-seq reads from NextSeq

[Maflores11]

Hi All,

I submitted 12 samples for ChIP-seq using the NexSeq Platform. Briefly, these samples represent 3 ChIP Biological Replicates:
Mock
Transcription Factor (TF)
Histone H3 (CT)
Input

According to the TapeStation, ChIPed DNA from Mock and TF were pretty low (500 pg total DNA), therefore, I decided to use Kapa Hyper Prep Kit. Barcodes (No.1 to 12) were from Bioo Scientific.
All samples were normalized to 500 pg initial DNA template, 19 PCR cycles.
After size selection (AMPure beads), samples were subjected to TapeStation. Average of DNA fragments ranged from 300-400 bp across all samples. We then were recommended the NextSeq platform, 75-bp, single end features. All samples were normalized to 6 nM and pooled in 20 microliters. We used the whole flow cell. After getting the raw data back, 11 samples contained poly AT-stretches, likely to be artifacts, only one Mock sample contained a low percentage of reads with those AT-stretches. Basically, 75% of the whole raw data contained such AT-stretches. I wonder if these artifacts were created during the library preparation or the NextSeq platform could introduce them. I'm attaching a PDF file showing the electropherogram of one sample and the AT-stretches present in our raw reads. Thanks a lot for the insight!

Miguel

[luc]

We are doing the library prep in a similar fashion (same kit and adapters) but are sequencing on the HS4000. We have not observed these poly-AT-stretches so far.

[Maflores11]

Dear Luc,

Thanks for the insight. I'm curious about the number of PCR cycles you guys are performing with your ChIP samples.
I used 75 nM of adapters for each ChIP sample when doing the library prep since higher adapter concentrations results in high adapter dimer amplification.
Would you be willing to recommend us a good Genome Sequencing Facility? We are considering to submit the rest of our ChIP samples to a different place.


Thanks a lot!

[luc]

14 PCR cycles ; our adapters are at 600nM.

[Lindeboom]

Dear Miguel,

We are having the same problem, and it is contaminating several libraries now that we are generating in a seemingly random fashion. We use the same workflow as you, apart from sequencing 50 bp, paired-end. We currently have no idea where this might be coming from, since samples prepared in parallel often result only several poly AT dominated libraries, making a contaminated stock-solution unlikely.

Did you in the meantime manage to figure out where your poly AT artifacts came from?

Kind regards,
Rik