Empty output from MACS2 when using control

Question

[lm687]

Hello,

I have been trying to perform peak calling with MACS2. I get an output when I don't use a control file, but blank files when I do. I'm reproducing someone else's results so there shouldn't be a problem with the input (I have only aligned the reads to the genome using BWA). I am running it with default parameters

macs2 callpeak -t SRR001992.bam -c SRR001999.bam

and changing the cutoff -q still yields blank files. Does anyone know where the problem could be? Thank you!

[rioualen]

Can you maybe post the output of macs2?

Default params include -mfold 5 50, try lowering it to 2 50 maybe. In several datasets I cannot get any peaks with the default mfold.

[lm687]

Thank you, I have modified the parameters but I still get the same blank output. The files peaks.narrowPeak and summits.bed are empty. peaks.xls is:

# This file is generated by MACS version 2.1.0.20151222
# Command line: callpeak -t bwaSRR001992_sorted.bam -c bwaSRR001998_sorted.bam -m 2 50 -n macs2_with_control/bwaSRR001992_sorted.bam
# ARGUMENTS LIST:
# name = macs2_with_control/bwaSRR001992_sorted.bam
# format = AUTO
# ChIP-seq file = ['bwaSRR001992_sorted.bam']
# control file = ['bwaSRR001998_sorted.bam']
# effective genome size = 2.70e+09
# band width = 300
# model fold = [2, 50]
# qvalue cutoff = 5.00e-02
# Larger dataset will be scaled towards smaller dataset.
# Range for calculating regional lambda is: 1000 bps and 10000 bps
# Broad region calling is off

# tag size is determined as 26 bps
# total tags in treatment: 4374226
# tags after filtering in treatment: 3824323
# maximum duplicate tags at the same position in treatment = 1
# Redundant rate in treatment: 0.13
# total tags in control: 4090488
# tags after filtering in control: 3608175
# maximum duplicate tags at the same position in control = 1
# Redundant rate in control: 0.12
# d = 131
# alternative fragment length(s) may be 131 bps

model.r contains vectors p, m, xcorr and x.

Thank you

[Sukhi Singh]

Another try would be to load these files in the browser and to see how similar the profiles are, that could be an another reason.

[lm687]

thank you, I will do this. Does the amount of total tags in the control seem reasonable?

[Sukhi Singh]

This I think has no limits, depending upon how much was sequenced and how the data was pre-processed. But macs integrates a scaling option, so for most cases, you don't need to worry about that.

# Larger dataset will be scaled towards smaller dataset.

[lm687]

thank you for the information!

[Sukhi Singh]

If you like someone's answer and they satisfy you, you thank people by upvoting them on Q/A platforms and forums.
Good Luck!

[xiang-jiao.yang]

How to solve this issue? I am having a similar problem.

Answer 1

[Sukhi Singh]

Another user had this problem, from the Tao's answer

Can’t see any problem. Some guesses:

1. qvalue cutoff is too high. Try lower it such as -q 0.1
2. disk full
3. chromosome names in ChIP file and control file are not consistent. e.g. ‘chr1’ in ChIP but ‘chromosome1’ in control

[lm687]

Thank you very much! Unfortunately, I don't think it can be either of these

Answer 2

[darsal2]

Did you ever figure out a solution? I am also having the same problem. All of my output files are empty... I'm running it using these parameters:

macs2 callpeak -t 4_D.bam -c 1_D.bam -g 2.5e8 -nomodel -n 4v1rep41_test -q 0.01

Answer 3

[macmath]

Could you check the parameter specifically based on the differences between regular peak vs broad peak

Example for regular peak calling:

macs2 callpeak -t ChIP.bam -c Control.bam -f BAM -g hs -n test -B -q 0.01

Example for broad peak calling:

macs2 callpeak -t ChIP.bam -c Control.bam --broad -g hs --broad-cutoff 0.1

Answer 4

[oriolebaltimore]

Any solution to this problem. I ran with IP and Input as control. I have 0 peaks list in xls file and also in narrowPeaks file.

Appreciate any solutions. I tried all changing q to 0.01, 0.1 , 0.5 etc. Also checked chromosome numbers formats etc.

Thanks Adrian

[goodez]

I think it's better to make this a reply or a separate question. But to answer your question I doubt there's much you can adjust to get more peaks. Sometimes there's just not much signal in the IP relative to control. Maybe the answer is to optimize the antibody or use another one.

[oriolebaltimore]

Thanks. I checked and I suspect the control - input FASTQ file is questionable. This was obtained from SRA.