coleslaw481
commented
2 years ago Sorry on the delayed reply. I was checking with the lab to see if anyone in our group was familiar with this codebase, but unfortunately the original developers left.
Couple niave questions:
- What does the distribution of values in the column for the search?
- Are the nodes in the huge module have a higher degree (more connected) then the other nodes?
If u are comfortable you could send your session file and what steps you ran and I could try to re-produce
NomiCentarix
Hello,
I have been trying JActiveModule with simulated annealing multiple times, with two big biological networks and p-values from several tests, and I also tried several combinations of the parameters.
However, in most of the cases, no matter what was the number of the modules or the iterations, I keep got the following pattern: one huge module, which encompasses of ~40-90% of the original network, and the rest modules are small ones, with single nodes.
Do you have any idea why I got such results?