I've used WGCNA to calculate adjacency and TOM for a set of 20k genes and particularly extracted adjacency and TOM for one gene relative to the 20k other genes. I would like to create a graph on cytoscape with the one gene as the central gene and the other genes to flank around it with respective to their adjacency or TOM with the gene.
Using the cytoscape app and importing the edges + nodes data (edge data loaded as network first and then node data loaded as a table into he network), I've been able to do this but I keep encountering an issue with accuracy of the mapping. I've used the edge weighted spring embedded layout but I get visualizations with some inaccuracy (i.e some genes with a higher adjacency may further away from the central gene than genes with lower adjacency even though the rest of the graph follows the intended trend).
There is a problem with randomness, when choosing the edge weighted spring embedded layout on the basis of adjacency as edge weight, the layout may randomly arrange genes with higher adjacency/TOM closer to the gene sometimes and vice versa other times . When I try to manually change settings in layout settings and apply the layout based on those settings, the applied layout will not reflect the selected column for "edge column that contains weights", the arrangement of genes around the central gene is random. I have colored nodes on the basis of pearson correlation and their shape height and width to be reflective of their average expression - if that has any impact on layout. All the weights (adjacency, TOM, etc.) have been loaded in as floats.
Would anyone happen to have any input regarding the randomness I'm seeing, it's unsettling that there are issues with accuracy and a lack of irreproducibility. I am using the latest cytoscape version (3.10.3). Thank you!
