A few questions about ATAC-seq library preparation. Thank you!

[JDHelix]

Hi everyone,

I've done my first ATAC-seq library preparation recently and had a few questions about the result. I would be very grateful if anyone could kindly give me some feedback/suggestions.

I used 50,000 HEK cells (fresh cells), and followed Buenrostro JD, et al. 2015 protocol exactly. After nuclei preparation -> Tn5 tagmentation (37C, 30min, on a heating block) -> PCR amplification (total cycle number was 11), I purified library with QIAGEN MinElute PCR Purification Kit and eluted in 20ul elution buffer provided by the kit. I loaded 10ul, 5ul, and 2.5ul samples on a 2% agarose gel (containing GelRed) to visualize the bands. Please see attached image.

I can see a nucleosome-like pattern at expected sizes, but not very strong. I'm wondering:
1. Is this result good enough for sequencing? Should I further optimize the condition (e.g. titrate Tn5 enzyme)?
2. There is a strong band below 100 bp (the lowest one), but I have no idea what it is. Could it be primer dimers? I have purified the library before running the gel, so I didn't expect to see primer dimers...
3. I expected to see strong signals below 150 bp, which come from sub-nucleosomal DNAs. However, I hardly see anything between 150 bp and the lowest band (indicated by a red square bracket in the image), even no background smear. Does anyone have similar experience and know why?

Thank you so much! Any comments would be highly appreciated!

Best,
Dan

[achouak]

hi,
i have a problem like you for atac seq. i am working in the same condition as the paper 2015 but i can not find the best results. i have many variations.
for some samples (fresh cells from mice) i can find many bands after run the gel but many others i can not i just find a large bands >1kb.
i tested 50000, 100000 and 300000 cells. i think it is better with 50000.
would you please tell what steps i have to change or to adapt to have more best results.
change our idea to ameliorate our results.
Achouak

[JDHelix]

Hi achouak,

I'm also a new user for ATAC-seq experiment.
Usually, ATAC-seq needs 500 - 50,000 cells. I think you probably have used too many cells in some cases. According to the paper, using too many cells causes underdigestion and creates high-molecular-weight fragments. It also makes sense that 50,000 cells worked better in your hand. If I were you, I would try fewer cells.
Good luck!

Best,
Dan

[María Camila Fetiva]

Hello,

Nowadays I am also having similar problems in my ATAC-seq protocol i don't have the banding pattern, could you finally resolve it and perform the ATAC-seq?.

Can you give me some suggestions?


Camila

[JDHelix]

Hi Camila,

First of all, I was wondering which protocol you were using? Howard Lab and Greenleaf Lab have published an improved ATAC-Seq protocol in 2017 (Corces, M. R. et al. An improved ATAC-seq protocol reduces background and enables interrogation of frozen tissues. Nat. Methods 6–8 (2017). doi:10.1038/nmeth.4396) which works much better than the original one. If you were still using their original ATAC-Seq protocol, I highly recommend that you try the new protocol.

Good luck,
Dan

[atpoint]

The original ATAC-seq protocol is very robust, and if you do it properly for human or murine cells you'll see the banding pattern. All these modifications in the newer protocols might improve the output but are not critical. If there is a problem, you'll not solve it by changing protocols. The only really essential addition from the newer protocols is the addition of Tween-20 at 0.1% to both lysis and transposition buffers as this really and notably decreased mitochondrial contamination from e.g. 80% to like 15% in our (and collaborators) hands. The other additions are rather minor in effect size. Camila, can you give details on your cells, the protocol and upload a bioanalyzer pic?

[María Camila Fetiva]

Hello this is the result I obtained and corresponds to sample number 1 please ignore the others.
I am following Buenrostro protocol in 50.000 cells. I did first 5 PCR cycles and next I added 20 PCR cycles.

I am not sure if this results is ok.

Thanks for your opinions and suggestions

[misgana]

Hi JDHelix,
I was wondering if you have taken into consideration that the adaptor sequences add extra 124 bp to you PCR product. May be that is why you don't see sub-nuclesomal size on your gel. Moreover, it will be wise to use 5% polyacrylamide gel rather than agarose for your electrophoresis.

[María Camila Fetiva]

thanks for your reply, do you think this Lybrary look ok? or there is not banding pattern, i am not sure if it is ok to sequence.

[narulam]

Hi Maria,

According to the protocol, you should do 5 cycles and an additional number of cycles based on the calculation from the qPCR done only on 5 uL of your PCR product for 20 cycles. If you're doing a total of 25 cycles for your entire DNA, you're probably over amplifying it which is not good for sequencing as it leads to bias toward few bp lengths. You should technically not amplify your DNA for more than 15 cycles total (including the initial 5 cycles). Hope this helps.

[SUKALPA]

This look fine.