Dear members of Biostar,

I am a Master's student doing a project in differential gene expression analysis using the Connectivity Map gene expression microarray data. The Connectivity map by Lamb et al 2006 has used 5 different cell lines treated with 1309 different FDA-approved drugs. I am only interested in looking at the MCF-7 (breast cancer epithelial) cell line and look at the differential expression analysis between the treated vs control samples of MCF-7 cell line. However, the problem is that per condition there is only 1 treated sample .CEL file and 5-6 vehicle control .CEL files, i.e. one .CEL file (treated with fulvestrant) vs 4-5 vehicle .CEL file (no treatment but only vehicle in DMSO).

Does this mean that there is no biological replicates in the Connectivity map? I have seen some papers who have performed differential gene expression analysis using the Connectivity Map. But I am not sure whether this is possible since biological replicate is not enough.

Please any thoughts or word of wisdom would be great

John from Cape Town