To obtain a high-level view of these Discovery compounds relative to known drugs, we ran t-SNE analysis on the Discovery signatures and those of every compound belonging to a PCL. t-SNE is a non-linear dimensionality reduction and visualization technique that attempts to preserve local-structure from high-dimensional datasets ensuring that samples that are similar in the high dimensional space are plotted close together in the embedding (Maaten and Hinton, 2008). t-SNE was run on consensus signatures across cell types for each perturbagen in landmark space, with initial dimensions set to 50 and a perplexity of 30.
