[NMF-user] cleaning dataset before running NMF

[ANNA ESTEVE CODINA]

Dear all,


I am running the NMF R package to classify cancer samples but I am not sure if if it is advisable to clean my gene expression dataset from unwanted variation before. For example, I have covariates like sex, level of degradation, etc...that are counfounders. If so, how should I clean it? Regressing out these factors and working with the residuals? Are tools like SVA or ComBat appropriate?


Thanks a lot for your feedback.


Anna Esteve Codina, PhD
Data Analyst
Functional Genomics Unit
Centre Nacional d'Ànalisi Genòmica
CNAG-CRG
Parc Científic de Barcelona -Torre I
Baldiri Reixac, 4
08028 Barcelona
email: anna.esteve at cnag.crg.eu
web: http://www.cnag.crg.eu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/nmf-user/attachments/20200701/0d5f9d38/attachment.html>