Figure 3From: Genome wide prediction of protein function via a generic knowledge discovery approach based on evidence integrationThe relative contribution of different evidence sources to prediction performance. This figure shows the contribution from individual data sets when input individually to the neural network. The plot axes are the same as Figure 2. The most reliable predictions come from MIPS data corresponding to physical interaction, genetic interaction and complex and gene expression profile data (cell-cycle data, the Rosetta Compendium, and MAPK data).Back to article page