Dear all,

I'm currently working on a ChIP-Seq for a transcription factor and was preparing libraries using the qiagen QIAseq Ultralow Input Library Kit, checking the ligation of adapters with the Kapa library quantification kit for Illumina sequencing and performing size selection with Ampure XP beads as described in the library kit protocol. The libraries were then checked on a bioanalyzer high sensitivity DNA ChIP.

The first replicate worked fine for me (some samples are included on the bioanalyzer Chips, sometimes the chip didn't run nicely, don't mind that), however, in the second replicate strange larger peaks appear in the bioanalyzer chip and also there are multiple peaks/bands instead of a normal distribution.

The shearing of both replicates was fine (done mechanically with a bioruptor) and targets could be recovered as checked in a ChIP-qPCR. Also, all samples that had been prepared in parallel looked weird (ChIPs and inputs), for both antibodies used and in replicate 1 for antibody 3, although all other samples were fine. If I purified those samples in parallel with a re-purification of an old sample. the old sample still looked fine, so I assume it has to do with the library preparation itself. Adapters were not reused and diluted to a 1:10 dilution as I was using 5 ng of input DNA. The number of cycles was chosen according to the Kapa library quantification and the minimum of cycles was used. For the PCR reaction I prepared a master mix and added it to the samples, so maybe that was a problem? Or do you think an enzyme could have gone bad? Is it just trash that got over-amplified, or could it be a problem with the adapter ligation? Or the most simple question: has anyone seen peaks like this before? And do you think it might be worth to give it a shot at sequencing nonetheless or is it just trash? For the libraries with antibody 1: could it be that those are "just" over-amplified? And if yes, could I try to sequence them, maybe after another round of size-selection to remove larger peaks?

The ChIPs were performed with primary cells (so I would be super happy, if I could use any of the samples) using two different conditions in wt and KO mice and the analysis therefore "just" should be the overlay between the different conditions (yes in wt, no in KO etc), only peak calling, nothing quantitative. Could I have a chance the sample quality might be sufficient for this setting? At least with antibody 1? Might the effect of this fragmentation into single peaks phenomena get lost once I pool the libraries for sequencing?

Thank you so much in advance.