Apply moderation to merged contrasts or to linear model only?

fgcz / prolfqua

Differential Expression Analysis tool box R lang package for omics data

https://pubs.acs.org/doi/pdf/10.1021/acs.jproteome.2c00441

MIT License

37 stars 7 forks source link

Apply moderation to merged contrasts or to linear model only? #82

Open Roman-Si opened 1 month ago

Roman-Si commented 1 month ago

Hi,

I have a question regarding when it is suggested to apply the Bayes moderation. In the vignette Comparing2Groups, the moderation is applied only to the linear model, and then the results of the ContrastsMissing are appended. In the publication of prolfqua the moderation is performed after merging the different contrasts. I go with the second method but wanted to confirm which method is recommended.

Thanks for your assistance!

wolski commented 1 month ago

Dear Roman,

Please stick with the paper's version —first merge and then moderate. I will update the Vignette to align it with the paper as soon as possible. Thank you for pointing out the discrepancy.

Best wishes