Hello everyone,
Could someone please provide a working example (or point me to a good resource) of how 'lambda local' (along with lambda BG, 1k , 5k and 10k) is calculated for peak calling in MACS?
λlocal=max (λBG, [λregion, λ1k], λ5k, λ10k)
Does 'max' here mean the upper limit, based on whichever of the variables λBG, λregion, λ1k, λ5k, λ10k is the highest and hence the p-value being calculated taking this 'max' value for λ? Is λlocal a product/average/sum of all the other λs?
I have been unsuccessful in trying to understand it by referring to the script and the tutorial given in the link below:
https://github.com/taoliu/MACS/wiki/Advanced:-Call-peaks-using-MACS2-subcommands#The_genome_background
Also, is the 'mfold' (10-30 fold enrichment) parameter estimated w.r.t lambda BG?
Thanks!
simple, assume you want to randomly pick a number between 1 to 100. If you do it many times (say 1000), how many times have you peaked a number between say 1-5?
(5/100)*1000 = 50
Where 5 is the interval length [1-5], 100 is the total numbers and 1000 is number of trials.
By analogy with MACS, interval length is estimated fragment length d, 100 is the estimated effective genome size and 1000 is the total number of reads form input sample, and 50 is λBG.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics