peak identification with macs2 for small RNA data

[skingan]

Hello,

I have illumina data from small RNAs and am using MACS2 to identify peaks in read mapping. (Read mapping done with bowtie1.) This method was developed for ChIP-Seq so I need to tweak the program settings for my purposes. My command is this:

macs2 callpeak -t mydata.bam -f BAM -g dm -s 29 --keep-dup all --nomodel --bw 15

I am using the nomodel setting to omit the peak shift model. Also set my tag length to 29 and bandwidth to 15. The peaks are still much wider than I expect, 200-300bp instead of 22-29bp.

Any thoughts on the program setting, or recommendations for a more appropriate method?

Thanks!