In the attempt on learning how to infer gene regulatory networks, I came across an infinity of methods and the data needed depend on the method (e.g ScenicPlus, scGPT, ...)
My question is very simple and theoretical, what data is needed to infer GRNs and could you provide examples with the matrix format (e.g scRNA-seq gene X cell)?
I know this is a very general question but I believe that many single-cell researchers may have the same question.
GRNs are typically directed graphs where nodes are genes and edges represent their regulatory interactions. Deciding on creating an edge between two nodes depends on the evidence for regulatory interaction you can get from the data at hand. Very often the evidence is weak and indirect so it's common to combine multiple data sources to try and refine the network. So what data is needed depends on what strength of evidence you want. The best data would be direct experimental evidence that gene X binds to the promoter of gene Y and does something to gene Y's expression.
Maybe this review can help.
