-
Notifications
You must be signed in to change notification settings - Fork 23
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Q: Peaks are too long and peak number is low #85
Comments
This seems like a threshold problem. It might be that for this dataset, the default thresholds for reporting peaks might be too stringent (for both HMMRATAC and MACS). Try lowering the threshold using the -threshold option. The current default is 30, but you might want to try 20 or 10. This should increase the number of peaks called. The model seems fine, as you pointed out, so you wont need to play with -u / -l anymore. |
Hi @WJH58, |
Describe the problem
Hi thanks for this tool!
I got very low number (~200) of peaks for each sample by hmmratac compared with the number of those peaks (~2000) called by macs2.
What solutions I tried
I tried more stringent -u and -l (e.g. l=25, u=40; l=25, u=50; l=25, u=60; l=25, u=100) however I still only get roughly 200 peaks. When -u exceeds 50, the training set will fail. Therefore, I sticked to l=25, u=40 and l=25, u=50.
This is one of the command line I tried:
This is one of my log file:
I think Baum-Welch algorithm worked because Aij are not 0.333, and signals of state 2 are higher than state 1. However, I only got 197 peaks (including 85 high coverage peaks). Some peaks are very long, consist of many small peaks.
This is one of the peak called by hmmratac:
Could you please help me solve this problem?
The text was updated successfully, but these errors were encountered: