Hi all,
I am using Next maSigPro package for my Next generation seq dataset (one condition and 4 Time points). Just a little confusion I have that need to resolve. Kindly help me out.
Suggested instructions from your paper and Manual, I have done my analysis but got not satisfactory results. The confusion is to provide a input matrix of Normalized counts or glmNB_counts matrix (obtained from edgeR by using glmFit and further glmLRT function). The problem is here that when I provide a TMM_counts matrix (cpm(y, normalized.lib.size=TRUE)) to maSigPro then I don't get any 1 significant gene while using glmNB_counts from edgeR (glmFit and glmLRT functions; obtained by glmNB=lrt$fitted.values), I got some significant clusters of genes. So, I really don't understand what's going on.
Could someone explain me why
Script: GLM_fitted_input_script.R ### gives some results
Script: maSigPro_TMM_script_no_results.R ### "no Significant genes"
Normalized counts:
> head(cpm_TMM)
WT.T1.1 WT.T1.2 WT.T1.3 WT.T2.1 ..................................
mmu-miR-28a-5p 69.0793247 47.624907 47.6001787 45.6811196 .................
mmu-miR-873a-5p 0.9729482 1.013296 0.5734961 0.9136224................
mmu-miR-466b-5p 0.9729482 1.013296 1.1469923 0.4568112.................
mmu-miR-3102-3p 419.3406891 484.355440 585.5395482 395.14168422.................
mmu-miR-5623-5p 0.9729482 1.013296 0.5734961 0.45681122.................
mmu-miR-410-5p 0.9729482 1.013296 0.5734961 0.45681122.................
> glmNB=lrt$fitted.values
> head(glmNB)
WT.T1.1 WT.T1.2 WT.T1.3 WT.T2.1 WT.T2.2
mmu-miR-28a-5p 56.1610380 53.9248040 95.278381 156.202447 149.752883
mmu-miR-873a-5p 0.8204116 0.7877443 1.391846 2.010414 1.927404
mmu-miR-466b-5p 1.0938815 1.0503250 1.855793 1.675347 1.606172
mmu-miR-3102-3p 510.3962471 490.0731642 865.897958 731.941241 701.719554
mmu-miR-5623-5p 0.8204116 0.7877443 1.391846 1.340277 1.284937
mmu-miR-410-5p 0.8204119 0.7877446 1.391846 1.340277 1.284938
I really have no idea whats going on
This is probably a better question for the bioconductor support site: https://support.bioconductor.org/
make sure when posting there that you also post your design matrix for edgeR and the relevant commands you ran (listing the files names is not that helpful)
What could be happening is that when you do pairwise comparisons between conditions there are significant genes, however, none of these significant genes remain significant when comparing between all time points (change in some but not all)