Geometric Escape Index as a Candidate Structural Descriptor of Viral Drug-Resistance Mutations: A Pilot Study with Two Independent External Validation Sets (n=97)
We propose the Geometric Escape Index (GEI = deltaR_LSP), derived from the Lo-Shu Projection (LSP) geometric framework, as a compact structural descriptor of mutation-induced drug-pocket deformation in viral proteins. Training set (n=21): SARS-CoV-2 Mpro, HIV-1 PR, Influenza A NA (rho=0.930, p<0.0001, AUC=1.000). External validation 1 (n=29): HIV-1 PR from Stanford HIVDB v10.2 (rho=0.823, p<0.0001). External validation 2 (n=47): Influenza A NA across H1N1/H3N2/H5N1/H9N2 (rho=0.933, p<0.0001, AUC=1.000). Combined (n=97): rho=0.890 (p<0.0001). GEI threshold 0.030 separated HIGH-resistance mutations across all datasets. Part of the Genesis Geometric Medicine Framework v4.
Authors
- Yao-Kai Kao
Institutions
- Allen Institute for Brain Science (US)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-05-11
- DOI
- https://doi.org/10.5281/zenodo.20127127
- Primary Topic
- Computational Drug Discovery Methods
- Type
- preprint