Entering edit mode

3.6 years ago

AP
▴
80

Hello everyone,

I have the list of differently expressed genes from DESeq2 for three different timepoints. I have few genes that are differentially expressed in all three timepoints for which I want to combine the p-values to get one. I was thinking of taking an average p-value, but looks like that is a wrong way. Also, is it ok to average the Log2FoldChange values of these genes too?

Can you suggest any r packages or methods to combine the p-values and Log2FoldChange values.

The three timepoints are from the same biological sample and same treatment condition.

Thank you, Ambika

What question are you trying to answer by doing this? There might be a more direct way to address it.

I am doing gene enrichment analysis of differentially expressed genes using TopGO for which I need p-value of the listed genes. In my case I have 3 different p-values from three timepoints. I am looking for a way to combine these p-values to one so that I can proceed further.

Can't you just take the maximum p-value (max_p) from these few genes and report it as p <= max_p. You can do the same with the foldchange.

The program I am going to use does the hierarchical distribution based on p-value so I don't think I can do that. I need an exact p-value.

Hi, Please, wich program you are using for hierarchical distribution? When I performed any kind of expression analysis (RNAseq or even microarray) with multiple factor (time, dose, tissue, etc) the programs usually have options to carry on with all the conditions OR you simply work with each dataset separately. I never heard anything about "combine p-values" in order to execute any downstream analysis. IMPORTANT: It is recommended to consider the

adjusted p-valuefor the significance of genes.