Benchmarked my gene co-expression analysis method against WGCNA. Both can detect early vs late development co-expression modules, but WGCNA returned more modules + higher spatial resolution (left). However, my method return tighter modules (right). github.com/cxli233/Simple… In heat map view, you can see WGCNA (left) is better at pinpointing tissue by developmental stage combinations.

Say goodbye to VennDiagram and UpsetR...Customized upset plots are pretty and clear. Upper left: total set sizes (geom_bar); upper right: legend (get_legend() from cowplot ; lower left: colored matrix showing intersections (geom_tile); lower right: intersection sizes (geom_bar). The code is now on github! From raw data to pretty plots... github.com/cxli233/custom… Fully customizable (colors, which subsets to plot) and extendable (adding new plots to the lower right end). Thanks everyone for sharing.