Hi everyone,

I am analysing cross-species transcriptomic data from mouse and human models treated with the same drug. The drug is known to act on a specific target gene, which I will call GeneX. My main goal is to assess whether the drug induces similar molecular responses in both models.

The mouse dataset is RNA-seq, while the human dataset is Agilent microarray. I am planning to compare differential expression results and pathway-level responses between species using orthologous genes.

I have two main questions:

Since the main goal is cross-species comparison, would it be better to filter the expression matrices at the beginning and keep only common mouse-human orthologs before performing differential expression analysis? Or is it preferable to perform the full analysis independently within each species and only filter to orthologs at the end?

The known target gene, GeneX, appears to be very lowly expressed in both models. In the mouse RNA-seq data, it is removed by filterByExpr, and in the human Agilent microarray data it is present but has very low signal intensity.

Given that the datasets come from different species and technologies, I know that direct comparison of RNA-seq CPM/logCPM values with microarray intensities is not appropriate. However, I would still like to show whether GeneX is detected or expressed at low/moderate levels in each model. Would you recommend any way to present this?

If anyone knows papers that address this type of analysis, I would really appreciate your suggestions.

Thank you!