Hi there!
I am analysing some transcriptomics data with the iPathwayGuide tool on advaitabio.com.
Turns our that a large number of differentially regulated genes in pathways that come up as significantly different between our test groups has very low expression counts, e.g. average count 15 in group 1, count 5 in group 2 and fold change 3 with significant p value.
Issue:
We question any real biological effect of genes with such low expression counts, and also question whether we are able to verify the expression of such genes with molecular biology methods.
Question:
We were advised to perform the pathway analysis with an input list that contains all genes, and then set a fold change and p value threshold. Now we are wondering whether we could pre-filter the list of genes (exclude genes with expression count < x) for pathway analysis. Would that be very bad for our statistics? How do people deal with this issue?
Many thanks for your responses in advance.
I am analysing some transcriptomics data with the iPathwayGuide tool on advaitabio.com.
Turns our that a large number of differentially regulated genes in pathways that come up as significantly different between our test groups has very low expression counts, e.g. average count 15 in group 1, count 5 in group 2 and fold change 3 with significant p value.
Issue:
We question any real biological effect of genes with such low expression counts, and also question whether we are able to verify the expression of such genes with molecular biology methods.
Question:
We were advised to perform the pathway analysis with an input list that contains all genes, and then set a fold change and p value threshold. Now we are wondering whether we could pre-filter the list of genes (exclude genes with expression count < x) for pathway analysis. Would that be very bad for our statistics? How do people deal with this issue?
Many thanks for your responses in advance.