Chapter IV - Supplementary files
Chapter IV Explainable AI for end-to-end target discovery and molecular design in plant pathogenic fungi Supplementary file S1 - Per-fold cross-validation metrics for APEX-Tar and APEX-Drug across all GNN architectures and baselines Supplementary file S2 - Cross-organism generalization of APEX-Drug: predictions on 228 experimentally validated druggable proteins from fungal and bacterial pathogens Supplementary file S3 - Explainability analysis of 38 validated Botrytis cinerea proteins: attention and GNNExplainer residue scores, insertion-deletion curves and perturbation metrics Supplementary file S4 - Proteome-wide dual-model predictions (APEX-Tar and APEX-Drug) Supplementary file S5 - De novo molecules generated by PMDM against the ADSL active site: 427 structures with docking scores Supplementary File S6 - Decoy-based enrichment analysis of PMDM-generated molecules versus physicochemically matched random compounds Supplementary File B1 - Training and cross-validation metrics for the bacterial APEX-Tar model on the VirulentHunter dataset Supplementary File B2 - Dual-model target prioritization across the Acinetobacter baumannii proteome: APEX-Tar virulence and APEX-Drug druggability scores