Entering edit mode
5.5 years ago
treywea
▴
10
Hi there, I have two different condition's histone modification's data without input control. Could I apply this software to call the true peak among them or I should sequence one more input control library? Looking forward to your comments!
Which software?
Yes you can use any software of your choice for comparative ChIPSeq analysis but, using input control have its own advantage like removing the bias occurring during the enrichment and sequencing process. i.e it is always needed to distinguish background noise from enriched signals.
Yes, thanks for your comments, I know that input control is important in chipseq analysis. But the situation is we don't have enough depth for that so we call peaks between two type of tissue directly. And we get some pre-result without including input-control. we are afraid that we could not publish the data without input control. Do you think some experiment like chip-pcr could help us for publication?
You should check your experiment on a genome browser. If the antibody is good, you will see clear and distinct peaks and little noise. If so, you probably will be fine even without input. In contrast, if you see signal all over the place, you might accumulate a lot of noise signal. Can you post a screenshot?
Towards the publication, some confirmation with ChIP-qPCR is always a good idea, as it not only increases the confidence of the data, but also shows that you put in work to hold up a good scientific standard. This will be a plus when it comes to reviewing your paper. Depending on the reviewer, some will not even notice that you do not have an input control. In my experience, reviewers often do not even look at the method section. It is most important that the results are well-described and well-presented (including a browser track screenshot), make biological sense and are supported by further experiments.
Thanks a lot for your insightful suggest! It helps a lot, I attached one of the DR peak I found. For most of the result shows a distinct and clear peak, and I set up a very strict cut off the detect the peak(qval<0.001 and fc > 3). I still get thousand peaks and the result is quite related to the gene I want to find. And we have 2 replicates. But the problem is how to convince my supervisor that the data without input control but with some experiment validation could get a so-so paper. I try to search some paper that this could work but nearly all the publication have input control now.
Thanks a lot for your insightful suggest! It helps a lot, I attached one of the DR peak I found. For most of the result shows a distinct and clear peak, and I set up a very strict cut off the detect the peak(qval<0.001 and fc > 3). I still get thousand peaks and the result is quite related to the gene I want to find. And we have 2 replicates. But the problem is how to convince my supervisor that the data without input control but with some experiment validation could get a so-so paper. I try to search some paper that this could work but nearly all the publication have input control now.
Just based on these two pictures, that looks fine to me. For histone modifications, tens of thousands of peaks are quiet normal in my experience. If the data make sense, I would probably go on with the project as planned. Please also see How To Add Images To A Biostars Post. You have to link the picture directly (so the URL ending with
.jpgor any other picture format). I did it for you this time ;-)thanks a lot! I will check that!