Pocketscribe: confidence-aware binding-pocket triage and molecular dynamics setup for predicted protein structures
Year: 2026
Type: software
Abstract: Pocketscribe takes a protein structure predicted by AlphaFold2, OpenFold, ColabFold, ESMFold or Boltz and produces a single self-contained report covering structure quality control, binding-pocket detection, druggability scoring caveated by the prediction's own per-residue confidence, optional cross-model pocket consensus, and ready-to-run GROMACS input files. Its distinguishing feature is treating per-residue prediction confidence as a first-class signal that caveats every downstream pocket claim, rather than analysing a predicted structure as if it were experimental. Its cross-model consensus module scores pocket agreement in an architecture-family-aware way, reporting agreement between independent method families separately from, and as stronger evidence than, agreement within one family. Research use only.
Author: Adedeji, John
Institution: Osun State University
Language: English
Subfield: Drug Discovery
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W7214353067