too much data in co-expression network drawing

Question

[melinaapashna]

Hi community i am having trouble drawing a co-expression network ... Actually i am working on endometrium’s expression data but the problem is that the number of the genes are too much even after all the normalization i knew to do on it and after the Ttest there is still about 3000000 (3 million) genes interaction ... i tried using scytoscape and there is a final output but the final network visualized is not clear ! I have tried using Cemitool but there is also the “no beta values found.it seems that the soft thresholding approach used by cemitool is not suitable for your data.” Error What can i do now ?

Answer 1

[scooter]

Well, 3 million edges are obviously not useful by themselves. I would certainly use Cytoscape's filters to choose a reasonable cut-off (if your data is normalized [0-1], 0.6 would be a reasonable starting point). One other approach would be to try to cluster the network using the coexpression as an edge weight. You will certainly need a lot of memory to cluster a 3 million edge network, though!

-- scooter