ATAC-seq small / overdigested fragments?

[Vgamma4]

Hi all,

First, I hope that the ATAC-Family will grow further to share their experiences here.
My experience with ATAC-sequence is kind of confuse. At the beginning I FACS sorted cells (20k - 50k) and followed the original Greenleaf paper. The fragments measured by BA were very small (one peak at 180bp). See Pic1

Then I simplified the process by just taking freshly isolated cells from lymph nodes, tried different cell numbers from (10k - 100k), with and without lysis (as decried in other threads. Suddenly I got this typically nucleosome pattern and run those samples at the HiSeq (see Pic2). The results were promising. mtDNA reads were without lysis at 10% and with lysis around 40%. 25k cells seemed also optimal for sequencing.

Then I went back with my optimised setting to the FACS sorting cells and got again only small fragments peak at 180bp. The sort by itself looked good.

Could it be that sorting cells influences the quality of the libraries meaning that the cells are stressed or maybe already dying. Any experience with small fragments? Could it be that the DNA is overdigested?

I am happy for any suggestions. Thanks

[mkel93]

Hello,
Did you ever optimise your protocol for sorted cells? I am having the exact same problem trying to do ATAC on 50K sorted haematopoietic cells (KSL).
Thanks

[Page]

Hey guys,

Have you already figured out the possible reasons? I also just get short fragments (around 200 bp, 400 bp), however, without any long fragments. At beginning, we also suspected that the DNA is overdigested and we did series of tests within gradient dilution of Tn5. However, it doesn't help. We still didn't get the pattern and we just have a 200 bp peak.
Any ideas? Do you think decreasing incubation time may help?

[xchrom]

Was this ever resolved? I’m having a similar issue with primary cells. Only finding ~200bp fragments and no nucleosomal banding pattern. Any thoughts would be appreciated!

[OlgaMikh]

Hello, has any of you solved this issue? Would greatly appreciate your response. I am seeing the same with FACS sorted cardiomyocyte nuclei.

[Meyana]

I have seen this on FACS'ed neuronal nuclei. Switching to the Omni-ATAC protocol (and handling the procedure a few times) solved everything and I now reliably get good looking BA tracks.
When removing buffer after sorting, take care to remove the whole buffer volume. Any dilution of the tagmentation reaction by residual sorting buffer clearly affected my early efforts.