[xenia.zhang]

captainentropy,

Thanks for your reply!

Actually I'm at the step before library construction. I figure I have to have decent amount of ChIP enriched DNA before I start doing library construction.

'half of the peaks are real' means, for example, I got 300 peaks from peak calling software (MACS/Cisgenome, etc). I manually checked these peaks and compared with other lab members' similar result and only about half of these peaks seem real. The other peaks are just background peaks that also exist in a control sample, or are too short to be real.

Hope this is clear?

thanks.


-Sherry

[hon]

Hi Sherry,

Have you figured your problem yet on low peak signal? Recently I have been in similar situations that couple of my histone ChIP and Anti-HA Tag ChIP are not working out quite well, and like you said resembling to those ChIP of inefficient antibody. I kind of in line with zhaoj who suggest that the ChIP DNA used for making library is not enough. Previously I chip-seq two samples, K4me3 and K27me3, both of them working out really fine, although I only use 1 millions cells for chip assay. I thought I could get similar success by using same number of cells for one transcriptional factor ChIP (HA tagged), and one histone ChIP (K79me2), however the sequencing results turned out are not good, e.g. not many peaks, and high FDR (30%). Similar to you, I am quite confident that the antibody should be ok, as I checked it by QPCR to show good enrichments on my known targets. While I am repeating the ChIP seq with more cells and waiting for the results, I wonder there could be other problems that cause the failure of ChIP seq?

For your bioanalyzer questions, I have tried the high sensitivity kit that can not detect the ChIP DNA, I guess because the quantity of ChIP DNA is really low, plus that they are in smear (not a single band), so even let's say if you have 5ng total but the smear may have 1000 bands, each band end up has 5pg, is already reach the limit of detection. In other aspect I guess all people using bioanalyser should pay attention is the bioanalyser is very sensitive to salt that means if there's minimal salt in your ChIP DNA that will ruin the measurement, I have seen this many times already, especially for phenol purified ChIP DNA which is not very clean and always contains a lot of salt. Hope this helps.

Hi all,

I have a couple chipseq libraries that show really good qPCR enrichment both before and after library construction, but sequencing results show very low signal, which resembles some other libraries where antibody is not optimal, though I know the Ab I use is a very good one. Then I start to suspect I have low ChIP enriched DNA for library construction, and tried to quantify concentration on high sensitivity bioanalyzer DNA kit. Unfortunately, for most of them, I can't detect anything, though the kit claims to be able to detect as low as 5pg/ul.

I wonder whether anyone else has this problem of quantifying chip enriched DNA with bioanalyzer high sensitivity dna kit, even with good Chiped dna samples? Then I wouldn't worry too much. Or it's really problem of my sample?

Or does any of you have any experience in this similar good qPCR, bad sequencing situation?
Million thanks to all!



-Sherry

[xenia.zhang]

Hi Hon,

Thanks for your suggestion. Current my explanation regarding your question is still not enough ChIP DNA. For histone modification, you don't need lots of starting cell since its so abundant and widespread. But for transcription factor, it's much weaker and rare. I know people in my lab who use half a million cell to do transcription factor chipseq. Although he did get result, total peak number compared with those using at least 20million cells is very small, and signal is not strong. Current my way is to use every chromatin prep for 2 ChIP, instead of 5 as before, and pool 2 chiped DNA together for library construction. I consulted experienced lab and they said that they've been successfully constructed library with 1-3ng of ChIPed DNA, but it's not always successful.

I suggest you can use picogreen (from invitrogen) to quantify your ChIPed DNA. I tried once, and although I haven't been able to get linear standard curve, my DNA is in the range, and after calculation, is around 3ng per ChIP (that's why I choose to combine two chips). The sequencing result seems to be fine now.

yah. That's my answer. Hope it will help!

good luck!
Sherry

Hi Sherry,

Have you figured your problem yet on low peak signal? Recently I have been in similar situations that couple of my histone ChIP and Anti-HA Tag ChIP are not working out quite well, and like you said resembling to those ChIP of inefficient antibody. I kind of in line with zhaoj who suggest that the ChIP DNA used for making library is not enough. Previously I chip-seq two samples, K4me3 and K27me3, both of them working out really fine, although I only use 1 millions cells for chip assay. I thought I could get similar success by using same number of cells for one transcriptional factor ChIP (HA tagged), and one histone ChIP (K79me2), however the sequencing results turned out are not good, e.g. not many peaks, and high FDR (30%). Similar to you, I am quite confident that the antibody should be ok, as I checked it by QPCR to show good enrichments on my known targets. While I am repeating the ChIP seq with more cells and waiting for the results, I wonder there could be other problems that cause the failure of ChIP seq?

For your bioanalyzer questions, I have tried the high sensitivity kit that can not detect the ChIP DNA, I guess because the quantity of ChIP DNA is really low, plus that they are in smear (not a single band), so even let's say if you have 5ng total but the smear may have 1000 bands, each band end up has 5pg, is already reach the limit of detection. In other aspect I guess all people using bioanalyser should pay attention is the bioanalyser is very sensitive to salt that means if there's minimal salt in your ChIP DNA that will ruin the measurement, I have seen this many times already, especially for phenol purified ChIP DNA which is not very clean and always contains a lot of salt. Hope this helps.

[hon]

So you said you combined 2 ChIP assay DNA to make library which works, do you know that equal to how many cells? Dose your transcriptional factor bind to a lot of genomic targets? I found if a TF doesn't have many targets may require even more cells. Right now I don't even bother to measure the ChIP DNA as I think that can not guarantee the ChIP seq will work. Even you ChIP with IgG which can give several ng of ChIP DNA that will make a good library but not give you any significant peaks. Seems people more incline to use more than 10 millions cells for TF ChIP, while histone ChIP can be in million range.

[saikumarkv]

Hi everyone,
In a previous experiment I used some MIRA enriched DNA (methylated) for a single read library prep (using Illumina Chip-seq kit) and sequencing, which worked very well. Right now I have lots of ChIP samples to sequence of various amounts (2-20ng). We decided we will use PE adapters for all sample preps from here on. I'm testing the NEBNext DNA sample prep kit and Illumina PE sample prep kits with different amounts of DNA (1-10ng). The question I have is whether anyone adopted the PE sample prep kit (where they use lot more reagents and enzymes) to the ChiIP-sample prep kit. Of course I will change the adapter and primer concentrations. I'm also testing TUFTS custom adapters and primers they they were gracious to share. Thanks in advance.