ChIP library prep results (Illumina)

[theduke]

Hi guys,

This is my first post here so thanks all in advance for any comments. I have been working on ChIP for a while now and am now just making the jump to ChIP-sequencing. Anyway, I have just prepared my complete post-PCR TF library and have attached the Bioanalyzer results (High Sensitivity DNA chip). I have a pretty nice peak around the 250 bp area, which is great, but as you can see there is second peak around 150 (adapters probably) and also some lower peaks which to me look like primer dimers.

My question is simply, how do you guys think it looks and how should I proceed? My immediate thought was to re-do the size selection on a much longer gel and simply cut out a fresh band in the 250 bp range. I have enough library to get away with this but I have no idea how this will impact sequencing downstream. Is it out of the question to proceed as is? Of course the other option is to re-do the whole thing, which would mean me going all the way back to the cell culture stage!

Basic info:

ChIP - 10 million cells, using Magnify ChIP (Invitrogen)
Illumina prep - as per Illumina protocol except with a 1:30 adapter dilution. Size selection was done first on 2% gel (mini-gel), followed by PCR (18 cycles)

[bfrost]

Hello,
Did you figure out what the smaller peaks were? I just made a library for the first time. I don't have primer dimers, and my sample itself looks good, but I have that same peak at 75bp that you do. Have you gotten rid of it?

[DMO]

I would definitely re-purify. What concentrations of adaptors and primers are you using?

[bfrost]

I figured out that that ~75bp peak is adapters (66 bases) plus primers (53 bases). You can just re-size select to get rid of it. The person at our sequencing facility said that my 75bp peak was less than 25% of my entire sample.

She said "if 25% loss is still sufficient for your needs (sufficient coverage) then I wouldn't worry." I don't totally understand what she means.

[NextGenSeq]

I would try to get rid of the primer artifacts, either run a gel or use Ampure XP beads.

[theduke]

I re-purified and that took care of everything and I still ended up with plenty to play with. I used 1:30 adapters and standard primer concentrations according to the Illumina kit. Following re-purifcation, I qRT-PCR confirmed the library and my sample is now running as we speak!

How long does a typical Illumina run take? I'm doing single-end, 40 cycles.

[NextGenSeq]

Read Length Run Time (Days) Output (Gb)
1 x 35 bp ~2 10 - 12
2 x 50 bp ~5 25 - 30
2 x 75 bp ~7 37.5 - 18
2 x 100 bp ~9.5 54 - 60
2 x 150 bp ~14 85 - 95
*Sequencing output generated using TruSeq SBS V5 kit with PhiX library and cluster densities between 508,000-630,000 clusters/mm2 that pass filtering on a GAIIx

http://www.illumina.com/systems/geno...lyzer_iix.ilmn

[josdegraaf]

@bfrost: If you would cluster with a library that looks like the one shown in the first post I would guess more then 30-40 % will be adaptors, those will cluster and be sequenced resulting on only 60 % real reads vs 40 % adaptors, so you loose 40 % real data.

[theduke]

Originally Posted by josdegraaf

@bfrost: If you would cluster with a library that looks like the one shown in the first post I would guess more then 30-40 % will be adaptors, those will cluster and be sequenced resulting on only 60 % real reads vs 40 % adaptors, so you loose 40 % real data.

OK so I have just recieved the initial results back form the sequencing team and I have attached the summary report. My samples are lane 7 (TF Chip) and lane 8 (input DNA). As you can see lane 7 gave very low % alignment (25%) where as the input looks pretty nice. I did not have any problems with the input library preparation. Whilst I do not know yet what the problem is with lane 7, I am willing to bet a lot of cash that it is an adapter/primer related problem in line with what josdegraf said.

What do you guys think? Will this data still be useable?

I will be re-doing the library with a modified gel purification procedure (longer, slower gel) and will probably lower the adapters even further (1:50 maybe). Has anyone lowered the primer concentration in addition to the adapter?

[cmawhinney]

Originally Posted by theduke

I will be re-doing the library with a modified gel purification procedure (longer, slower gel) and will probably lower the adapters even further (1:50 maybe). Has anyone lowered the primer concentration in addition to the adapter?

I try and dilute my primers a bit, especially if I don't have the recommended 10ng of ChIP DNA to play with, ie, if I have 2ng of DNA then I will likely dilute my primers 1:5. I then clean up my PCR amplification using SPRI beads as opposed to a Qiagen column. I never had an issue with adapter contamination (since I cut my gels at around 250bp) but I used to have an issue with primer dimers. SPRI beads solved that issue for me.

Do be careful with your primer dilutions though. If you dilute too much you'll wind up having too much template and not enough primer and your template will start annealing to itself.