search

matthiasbock / BooleSim

Systems biology tool: Boolean network simulator
http://matthiasbock.github.io/BooleSim
11 stars 6 forks source link

Importing from GINML adds wrong nodes #38

Closed matthiasbock closed 11 years ago

matthiasbock commented 11 years ago

Import from GINML doesn't work properly. Additional nodes with name "FALSE" are added in JavaScript. They do not show up in the network. Why are they there ?

These nodes are getting exported to Python and R aswell, which breaks the model.

chaitan3 commented 11 years ago

https://github.com/matthiasbock/BooleSim/issues/28

Refer to the last 2 comments

TillS commented 11 years ago

Yes, we know, but we re-opened this issue for a reason. Since we're about to re-submit in a few days, we really need this issue to be sorted out, because although our tool doesn't interpret the wrong falses, they still appear in python and R export and will make trouble if interpreted with other software. If someone for instance wants to extend an existing .ginml model with our software then analyse the extended model with a package like BoolNet, he will encounter trouble! This is a reason for rejection when I look at the previous rejection reasons of the reviewer and we don't want to be rejected again I think. So shouldn't it be possible to erase the remaining falses? It would definitely be great if we could do that and I think this will really be the last remaining main issue to fix now! cheers!

chaitan3 commented 11 years ago

Ok, agreed.

The reason I was apprehensive about fixing the issue is because I did not want to tamper with the GINML parsing code. During GSoC I remember that I had spent a lot of time debugging the simulation of GINML files. One small error in the parsing code could alter the entire state space. But as you have pointed out that it may be a point for rejection, the bug should be squashed.

PS: Matthias's latest commit should fix the issue I think.