Entering edit mode
9.0 years ago
s.vdrzeeuw ▴ 20
I am just started working with MACS and noticed that it calculates a pvalue per peak. But where does it come from and how is it calculated.
Have checked several forums but unfortunately couldn't find any explanation on how MACS does this. Anyone a good example on how MACS calculates its pvalues?
Read the paper: Model-based analysis of ChIP-Seq (MACS) Zhang et al., 2008. ( http://www.ncbi.nlm.nih.gov/pubmed/18798982) There's a clear explanation about dividing the genome up into bins and using the Poisson distribution to estimate the P-value for the number of reads observed in any particular location.
Can you explain why macs14 by default chose so low pvalue (0.00001)? Can I chose it 0.05 in case I am finding only peaks with this?
You can choose whatever P-value you want, there's nothing magic about 0.05. However, depending on what you're studying, if you look at typical MACS p-values for "good" robust peaks they are often many many many orders of magnitude lower than even 0.00001. So make sure to look at the tracks and convince yourself of the relationship between what MACS assigns as a peak call with a P-value, and what you see.