Dump norm KR failed for certain resolution

249 views
Skip to first unread message

Hurley Li

unread,
Aug 3, 2016, 3:07:34 PM8/3/16
to 3D Genomics
Hi, 

When I run juicebox dump norm KR on a .hic file, sometimes it failed at certain chromosome, but only at some resolution but not others. 
For instance, I have this error "Norm not available at chr14 1700 BP KR", also not available at 2000 BP, but I do successfully dumped for resolutions like 1100BP, 1200BP, 1300BP... etc. I'm not intend to use these weird resolutions, these are just for testing, but I do hope I can get norm KR at 2000BP. 
I know if the matrix is too sparse, I won't get KR at certain bin, but my data is at resolution ~2kb. Is there any specific reason that the norm KR was not available at certain resolution? My juicebox pre did not generate any error. 

Thanks in advance!

Hurley

Neva Durand

unread,
Aug 3, 2016, 4:16:50 PM8/3/16
to Hurley Li, 3D Genomics
Hi Hurley,

The norm KR is available when the KR matrix balancing algorithm converges.  When there's no convergence, sparse rows are successively thrown out and the normalization is tried again, up to a certain point; in the final iteration, the 10% sparsest rows are thrown away and if convergence still fails, a vector of NaNs is returned.

Since it seems clear that you are comfortable making your own jar, you might want to modify how many times the normalization is attempted before it fails.  Have a look at the file juicebox.tools.utils.original.NormalizationCalculations.java, and look at the function computeKR.  You would modify the code around "iteration" to be something higher.

After modifying, instead of calling juicebox pre, you should call juicebox addNorm <hic file> with your previously created hic file.

Best
Neva

--
You received this message because you are subscribed to the Google Groups "3D Genomics" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3d-genomics...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/3d-genomics/5b3c252f-746a-4d1c-ba43-676aebb055c7%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
Neva Cherniavsky Durand, Ph.D.
Staff Scientist, Aiden Lab
Reply all
Reply to author
Forward
0 new messages