Biological significance of Fold_enrichment in macs2

Question

[Rose]

Can anyone explain me about the biological significance of fold_enrichment in macs2.

[ATpoint]

In order to motivate people to answer, provide details on what you are working on, what you have done so far and what your view on the matter is, as Devon already suggested.

[Devon Ryan]

What do you think it means?

Answer 1

[Ian]

The fold enrichment score compares the read/fragment count at the binding region summit against the local background model. For me it is the second most useful statistic after the qvalue. After sorting by qvalue I sort by fold enrichment. I tend to discount regions with FE < 5, as these tend to represent results with a high genomic background. This cut off is flexible based on known binding regions, for example. I am not sure whether it has biological significance, more a way of selecting a high confidence set of results for downstream analysis.

[Rose]

Thanks Ian, I have one more doubt. I have run chip-seq experiments on mouse samples and did peak calling using macs2 by the following command. pol1_1.fastq is the control sample and pu1_1.fastq is the chip sample.

macs2 callpeak -t pu1_1.fastq_trim.gz_A1.bam -c pol1_1.fastq_trim.gz_M1.bam -n pu1_1.fastq_trim.gz_A1.bam_macs2 -f BAM -g mm -B -q 0.01 --nomodel

It generated pu1_1.fastq_trim.gz_A1.bam_macs2.xls file. but all the fold enrichment values were positive. I couldnt find any negative values. what is the reason?

[Devon Ryan]

Negative enrichments aren't ChIP peaks, so they won't be returned.

[Rose]

Thanks Devon! So, In case, if I need the negative enrichments, I should swap the control and chip sample files while doing peak calling?

[Devon Ryan]

Yes, since negative enrichments are also called "peaks in the control".

[Rose]

Thank you so much Devon

[Ian]

Sorry I missed your question, but Devon was absolutely right.