Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Vinícius A. Paiva¹, Douglas E. V. Pires², Gustavo C. Bressan¹, Sandro C. Izidoro³, Sabrina A. Silveira¹ (1. Universidade Federal de Viçosa, Brazil | 2. University of Melbourne, Australia | 3. Universidade Federal de Itajubá, Brazil)
BENDER DB is a comprehensive database for exploring predicted protein binding sites across the proteomes of pathogens associated with neglected diseases. Using AlphaFold-predicted structures, the database integrates predictions from five distinct tools—GRaSP, PUResNet, DeepPocket, PointSite, and P2Rank—covering over 100,000 proteins and totaling over one million predicted sites. In addition to individual outputs, BENDER DB offers consensus analysis to highlight high-confidence residues predicted by multiple methods. The platform also introduces BENDER AI, a supervised learning meta-predictor that combines the results of the five tools, improving accuracy and reducing false positives. BENDER AI achieved superior performance (MCC = 0.64, AUC = 0.89) compared to individual predictors. BENDER DB provides a unified interface with molecular visualization, searchable tables, and comparative tools, supporting target identification and drug discovery efforts for neglected diseases. The database is freely available at: https://benderdb.ufv.br.