Question: Differential expression of GSVA Enrichment Scores - voomWithQualityWeights and lmfit
gravatar for SB
6 months ago by
United States
SB0 wrote:

Dear Bioconductors,

I am working on using GSVA to look at differential expression of gene set enrichment scores across several sample groups and the dataset I'm using is from NanoString nCounter (which can be treated like RNA-seq count data). I have read that it is recommended to normalize the data with voomWithQualityWeights prior to using linear modeling with limma. However, since I am planning to perform GSVA and then limma for differential expression of the scores, would it still be appropriate to use voomWithQualityWeights?

My next question is about using lmfit with GSVA enrichment scores. In the documentation and many examples I have seen online, I haven't seen use of the block argument or estimation of the correlation value for differential gene set analysis. Is it not recommended to utilize the block or correlation argument in lmfit at the set level?



ADD COMMENTlink modified 22 days ago • written 6 months ago by SB0
Answer: Differential Expression of GSVA Enrichment Scores - pre-processing with voomWith
gravatar for Robert Castelo
6 months ago by
Robert Castelo2.3k
Spain/Barcelona/Universitat Pompeu Fabra
Robert Castelo2.3k wrote:


others may give you more specific advice about using voomWithQualityWeights but if you look at the documentation you'll see that this function allows you to use the power of voom with the power of sample-specific weights. The latter are specially useful when samples have very different qualities and you do not want to discard any of them. That said, while voom (and voomWithQualityWeights) has a 'normalize.method' argument, the method itself is not a normalization method. I'm saying this because the best way to normalize NanoString nCounter data is not the same question as wheter voom is suitable for that kind of data, which I don't know.

regarding the normalization of GSVA enrichment scores there has been already discussions on this topic with RNA-seq data:

Pre-processing RNA-seq data (normalization and transformation) for GSVA

I do not know the specific features of NanoString nCounter data. If you think that you need sample-specific weights to do your differential expression analysis at gene set level with  GSVA enrichment scores, one thing to try would be to transform the NanoString nCounter data to log CPM units, using the 'cpm' function from edgeR for instance, and then use the function 'arrayWeights' from the limma package.



ADD COMMENTlink written 6 months ago by Robert Castelo2.3k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 250 users visited in the last hour