brianjohnhaas
commented
1 year ago Hi Dario,
I think it's just the HMM and Bayesian predictions that require counts for modeling. If you need to feed counts back into infercnv, maybe you could take your batch corrected data and scale it such that the column sum approximately matches the original column sum in the pre-batch-corrected data.
On Sat, Feb 25, 2023 at 4:00 PM Dario Strbenac @.***> wrote:
The requirement of count data seems like a tough limitation. I would like to use batch effect correction methods, some which produce decimal data that no longer has count chracteristic. Then I would like to run inferCNV on batch effect corrected data. I have matched whole genome sequencing on these cancers and I want to know if the batch effect correction is destroying true biology by comparing whole genome sequencing arm losses and gains to (1) inferCNV on raw count data and (2) on batch effect corrected data. If batch effect correction is working well, the arm losses and gains identified by whole genome sequencing will be preserved in (1) and (2).
— Reply to this email directly, view it on GitHub https://github.com/broadinstitute/infercnv/issues/515, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABZRKXYKEWPILDT2EHH45XDWZJXF5ANCNFSM6AAAAAAVIBZLUI . You are receiving this because you are subscribed to this thread.Message ID: @.***>
--
Brian J. Haas The Broad Institute http://broadinstitute.org/~bhaas http://broad.mit.edu/~bhaas
DarioS
The requirement of count data seems like a tough limitation. I would like to use batch effect correction methods, some which produce decimal data that no longer has count chracteristic. Then I would like to run inferCNV on batch effect corrected data. I have matched whole genome sequencing on these cancers and I want to know if the batch effect correction is destroying true biology by comparing whole genome sequencing arm losses and gains to (1) inferCNV on raw count data and (2) on batch effect corrected data. If batch effect correction is working well, the arm losses and gains identified by whole genome sequencing will be preserved in (1) and (2).