Hi, I am new to ChIP-Seq data analysis. In my samples, the expression of transcription factor of interest was observed to be down-regulated. Subsequently, a ChIP-Seq on this transcription factor of interest in three replicates was performed and input DNA was used as control. Initially when I performed MACS2 analysis using the command below:
mac2 callpeak -t treatment.rep1.bam treatment.rep2.bam treatmen.rep3.bam -c input.rep1.bam input.rep2.bam input.rep3.bam -f BAM -g hs -n output.macs2
The run failed and I received the error below:
WARNING @ Wed, 10 Oct 2018 11:53:34: Too few paired peaks (0) so I can not build the model! Broader your MFOLD range parameter may erase this error. If it still can't build the model, we suggest to use --nomodel and --extsize 147 or other fixed number instead.
After adding --nomodel to the command. The run was successful but not peak was outputted.
My question is: Is it expected for no peak to be detected for samples with down-regulated TF? Or should I further optimize the parameters to increase the chance to get some peaks?
I was wondering if there might be at least some peaks detected due to noise.
Thanks in advance.
If you heavily knock-down TF expression it's unsurprising that there are no peaks. I wonder, how well the library prep worked. Were you able to get decent amounts of DNA from the IP still? Have you looked at a fingerprint from
plotFingerprint? If you got a decent amount of DNA still and the fingerprint looks OK then often playing around with--slocaland--llocalhelps.Hi,
You should maybe start with generating coverage files (wig, tdf, bedgraph...) and checking what your samples look like with a genome browser, see if it looks alright.
As for Macs2, as it says in the error message:
Broader your MFOLD range parameter may erase this error.By default, the mfold range is 5-50, maybe you can lower the threshold and see what happens:Can you be sure that the antibody is of high quality and able to do decent IP if true peaks were present.