Dear List,
I have done time-course experiment with two data series (control and cold compared to time-zero reference sample -- so it's log2ratio data). I have two lists of differentially expressed genes over time (from maSigPro package):
- genes with significant changes in control
- genes with significant changes in control and cold
For each series I have seven time points. I'm interested in changes over time in cold treatment but I'd like accout for changes in control.
I'd like to know your opinion -- how would you cluster genes significant in both cold and control?. I've tried to cluster merged data (i.e. table ca 600 rows (genes) times 14 samples >>seven cold points + seven control points<<). I've done clustering in Mfuzz package and I'm not fully satisfied with results.
Maybe I should cluster profiles which reflects differences (log-ratio "Cold/time zero" minus log-ratio "Control/time zero" for each time point)? Personally I don't like this idea because it looses information.
Or try to make some biclustering?
