Is it possible to add differential network analysis function in the package in the future?

[BW15061999]

Hi, Thanks for providing such a great package! And do you have a plan to add that function in this package? Differential network analysis can be used to identify changes in connectivity patterns or module structure between different conditions

Best!

[smorabit]

Hi, yes this is possible but I don't currently have any timeline for adding this into hdWGCNA. The TOM representation of the network can be extracted using GetTOM if you would like to perform a custom analysis. There are so many different ways of comparing networks, did you have a specific comparison/test that you had in mind?

[BW15061999]

Hi, if the differential network anaylsis is based on WGCNA, DiffCoEX may be a good choice. This is the main part of DiffCoEX, and its downstream is quite the same as orinigal WGCNA.

beta1=6 #soft thresholding
AdjMatC1<-sign(cor(datC1,method="spearman"))*(cor(datC1,method="spearman"))^2
AdjMatC2<-sign(cor(datC2,method="spearman"))*(cor(datC2,method="spearman"))^2
diag(AdjMatC1)<-0
diag(AdjMatC2)<-0

dissTOMC1C2=TOMdist((abs(AdjMatC1-AdjMatC2)/2)^(beta1/2)
                    ,TOMType = "unsigned"
                    )

[wendelljpereira]

Hi @smorabit,

Thank you for the great tool. I have been using it for the past couple of months, and it is excellent!

The only feature I missed was a way to compare different networks, as mentioned by @BW15061999. Experimental designs where there is a control dataset vs. one or multiple treatment datasets are very common. I believe that being able to separate connections that are only meaningful in the treatment would be very useful.

In my case, I have multiple time points that I would like to compare with the control (0h). I was able to find modules that correlate with the time points, and they are interesting, but a more refined way to select modules of genes (or even significant co-expressions among genes) that are specifically present in the more advanced time points after the treatment would be helpful.

I am unsure how to achieve this, but I will extract the TOM matrices and look into the literature for methods to compare networks. During a quick search, I found this example of a tool that does something similar to what I would like to be able to extract from the data: https://seidr.readthedocs.io/en/devel/source/compare/compare.html.

Best Regards, Wendell