PALMER: improving pathway annotation based on the biomedical literature mining with a constrained latent block model

摘要

Performance evaluation of PALMER using synthetic data, where a–d show the two scenarios we considered in our simulation studies. Each heatmap shows the binary matrix, where a black cell indicates value of one. In each heatmap, the color bars on the left show the true [left: “Gene (true)”] and the predicted gene clusters [right: “Gene (pred.)”], respectively. The color bar on the top indicates the predicted GO term clusters [“GO (pred.)”]. Error rates for gene and GO term assignments are provided