Suppose in an experiment I have three biological replicates of a single condition. In looking at the data on MDS plot, I can see that one of the three replicates appears far away from the other two replicates (which are grouped tightly). Is there a way I can identify which genes are driving this clustering? There are several features in cummeRbund that allow you to identify patterns between conditions, but how to look at a single replicate isn't obvious to me.
The motivation is to try and figure out what to change in future experiments. Stringent randomization was taken at all steps from performing the experiment to sequencing, and a great amount of effort was taken to ensure uniform handling of all samples, but the clustering analysis is showing some significant problems. I'd like to determine what may have caused the large distance among replicates in this condition by looking at what differences in expression there are. Depending on what the genes are involved in, it may give us some insight into what happened.
Any suggestions?
