Hi,
I am running EdgeR as a part of a bigger pipeline (CAP-miRSeq to be precise) where I quantify the expression and differential expression of miRNAs. However, the values in differential_expression tables does not match (though they are usually close) with values in expression tables (miRNA_expression_norm.xls, mature_miRNA_expression.xls and miRNA_expression_raw.xls). I am not sure which read count value were used in the differential expression tables.
Does EdgeR also calculate the read count first before doing differential expression calculations? There is another tool in the big pipeline which has the capabilities to calculate read count. But I am not sure if the read counts come from that tool or from EdgeR.
Can you please tell me where what could be the reason for different numbers in expression and differential expression tables?
•
written
by
Floydian_slip
EdgeR take read count as an input instead of performing the counting itself. What EdgeR will do is that given the read count, it will calculate the CPM for the statistical analysis. (Thanks for Devon's correction, EdgeR will not use CPM for statistic calculation) Maybe the value you see in the EdgeR output is the CPM instead of the raw input?
The differential expression (topTags) tables from edgeR don't contain expression values, so it would seem that you are comparing different things. The topTags tables do contain an Average Log CPM value, but this is an average for the miR across all samples, not an individual expression value.
edgeR analyses the counts but doesn't compute them. Many people use subread-featureCounts or htseq to compute read counts.
•
written
by
Gordon Smyth
Log in to answer this question.
