Thank you very much Carlo. I will check DEqMS package.
Hi,
I am new to BioInformatics and I would really appreciate if you could help me with this problem. I have a proteomics dataset with the following columns:
Protein CTR_1 CTR_1 T_1 T_1 CTR_2 CTR_2 T_2 T_2 CTR_3 CTR_3 T_3 T_3
Basically, for each protein, I have pairs of control and treatment samples collected at three time intervals (0h, 12h, 24h).
I want to find a p-value for proteins to identify differentially expressed proteins.
Could you please help me decide which test(s) I should perform here? Should I do three different two sample t-tests (if so, how would I get a single p-value for each protein?) or should I perform a repeated measure anova or something else?
Also can I use limma package to do so as I believe it provides ones with better p-values compared to using t.test( ).
Thank you!
2 answers
You probably should use a dedicated tool for proteomic analysis. For instance the R package DEqMS is especially designed for that purpose. It implements linear modeling (an extension of ANOVA) with some improvements to accommodate for the specificity of spectral count data.The documentation explains how build a factorial design and compute the pvalues associated with the factor effects (treatment and time in your case).
Hi, Are they replicates or repeated measurements? Also, I am aware of medical proteomics (technical/biological) while and some plant proteomics folks handle these terms differently.
To have an idea please refer this https://www.reneshbedre.com/blog/repeated-measure-anova.html
This is a good place to refer https://bioconductor.org/packages/release/bioc/vignettes/DEP/inst/doc/DEP.html
Also here authors have mentioned useful packages for proteomics https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8283032/
Thank you for the references. They are repeated experiments at time 0h, 12h and 24h time gaps.
Log in to answer this question.
Hi, Appreciate if you could point me to a reference book that I can follow in this regard.