Although biological pathways are essential forinterpreting results from computational biologystudies, the growing number of pathway databasesmakes it difficult to perform pathway analysis.Our study seeks to reconcile pathways from dif-ferent databases and reduce pathway redundancyby revealing informative groups with distinct bio-logical functions. Uniquely applying the Louvaincommunity detection algorithm to a network of4,847 pathways from KEGG, REACTOME andGene Ontology databases, we identify 35 distinctcommunities of pathways and show that thesecommunities are consistent with expert-curatedpathway categories. Further, we develop an algo-rithm to automatically annotate each communitybased on member pathways’ names. By learn-ing informative categories, we progress towards atool that computational biologists can use to moreefficiently interpret their biological findings.