Normalize Expression Matrix

[Fernas]

I have expression matrix where rows represent samples and columns represent time points. and values represent absolute counts of expression. I notice that some of the samples are highly expressed in all time points while some are less expressed in all time points. For examples:

Sample1: 1000 6134 6245 3341 ..etc
Sample2 2 3 4 2 ..etc

So plotting heatmap for such matrix and/or finding differential expression will not be possible.

So, I think I need to normalize the matrix and/or rescale the matrix to a common library (maybe to remove the sequencing bias!!). Am I correct?If yes, is there any known method(s) for doing that? Then, is there a library in R for doing such normalization?

[dpryan]

There are a LOT of options for this. If this is RNAseq, then just use one of the standard packages. If not, many of the microarray packages (particularly limma and affy) will have a large variety of options.

[Fernas]

Thanks @dpryan.
Actually, the counts are from metagenomic data from different annotation categories, taxonomy, kegg. So, I am wondering which normalization technique would be suitable?