Journal of Medical Sciences and Health Care Research
Desmoplasia-Aware h-Foci Expose Repurposing Limits to Prioritize Targeted Degradation and Cell-Therapy Anchors in Pancreatic Cancer
Abstract
Fred Y. Ye
Expression-signature reversal and molecular docking can generate apparently positive repurposing claims without testing therapeutic validity. We analysed 353 pancreatic tissues, a continuous-h differential network, 48,570 tumour- atlas cells and KRAS-mutant LINCS perturbations. Of 14,954 genes, 4,048 were concordant; a 295-node backbone contained four cell-localized h-foci and 61 second-order hubs. Seven of 541 active-FDA perturbagens passed repeated reversal, but all failed mechanism or clinical gates; metformin reversed nine of 12 signatures despite two negative randomized trials. We therefore routed the 61 hubs by compartment and disease direction: 9 were cell-surface, 9 extracellular/secreted, 16 internal-membrane/organelle and 27 soluble/nuclear. CDH3 emerged as a clinically calibrated surface anchor, while GPRC5A and LAMC2 were high-contrast non-hub backbone nodes. LAMC2 requires extracellular lysosomal targeting rather than canonical PROTAC; SAMD9L and XAF1 received degradation vetoes. This falsification-to- modality framework yields testable cell-therapy, antibody and degradation hypotheses without converting computational proximity into efficacy claims.