Question about createFitHiCFragments-fixedsize.py

53 views
Skip to first unread message

kaivalya shevade

unread,
Feb 27, 2023, 8:18:23 PM2/27/23
to Fit-Hi-C
Hi,

I know the format of the fragments input file should be as shown in the readme page
chr extraField fragmentMid marginalizedContactCount mappable? (0/1)
1 0 15000 234 1
To get the marginalizedContactCount I assume it should take a contact matrix as an input.

However the only inputs required to run this script are chromosome lengths and resolution. And the output it produces looks like a bed file which looks like this:

chr1 0 2500 1 1

chr1 5000 7500 1 1

chr1 10000 12500 1 1

chr1 15000 17500 1 1

Am I missing something ? What do you think I should do to get the output in the correct format ?

Thanks,

Kaivalya


Ferhat Ay

unread,
Feb 28, 2023, 12:34:31 AM2/28/23
to Fit-Hi-C
Hi,
Both are the correct format, the two fields that matter the most are 1st and 3rd and in both cases they list the "chr" and the "midpoint", respectively. 
For fields 4 and 5, they are only relevant for filtering out of "bad" fragments with low mappability or marginalized counts etc. However, that filtering will practically be achieved when you use a bias file so the fragments file for fixed-size binning case ended up pretty useless for the case when you use fithic with normalization biases as input - which is what you should do anyway. 
Hope that helps

Reply all
Reply to author
Forward
0 new messages