AI Science Tracker
EN

Learned screening identifies early antiviral candidates with enzyme and cell-entry assay activity

National Center for Advancing Translational Sciences, National Institutes of Health; Texas Biomedical Research Institute; National Institute of Diabetes and Digestive and Kidney Diseases · 20 Mar 2026 · biology

Discovery

Learned screening identifies early antiviral candidates with enzyme and cell-entry assay activity

Source check: Source checked. Not independent scientific confirmation.

Finding: A two-stage learned screen combined SchNet molecular representations with XGBoost to prioritize compounds for laboratory evaluation. Follow-up assays identified early candidates with protease-inhibition and cell-entry activity, with additional cell and human-airway tissue-model tests. These are moderate-potency laboratory hits; direct target binding, clinical benefit and reduced resistance were not established.

Technology: SchNet molecular graph classification; XGBoost candidate prioritization

Publication: Peer-reviewed paper · Test context: direct test. Catalog categories, not an evidence grade.

AI contribution
SchNet learns molecular representations from prior screening measurements and filters library compounds; a second learned XGBoost layer incorporates descriptors to prioritize candidates before new laboratory measurements.
Test and result
The learned screen prioritized roughly 500 compounds from about 170000 library entries. The source reports 27 initial follow-up hits and examines a smaller representative subset using enzyme, cell-entry, cell-based and 3D airway-tissue assays. Four of eight detailed representatives showed moderate protease inhibition; others primarily showed entry-assay activity.
Limitation / still unknown
This is not evidence that every hit has a confirmed dual mechanism. Direct specific target binding remains unconfirmed; aggregation or protein-destabilization artifacts were not excluded. Some tissue-model effects coincide with cytotoxicity, and incomplete inhibition complicates potency estimates. Reduced resistance, patient efficacy and a clinical therapeutic window were not tested. The current readable journal version is represented; the publisher version-history endpoint was rate-limited, so exact timing of any later text updates is unresolved.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://www.mdpi.com/1420-3049/31/6/1043

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Exact primary front DOI/title/date and fresh primary notice metadata; computational model/data Methods and high-level assay results/limitations read. No operational pathogen procedures represented or archived in the case evidence..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 3/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse biology discoveries.