SMA Research Platform

Screening Hits

COMPUTATIONAL

Positive binding predictions from AI-driven virtual screening. Each hit is tracked through the v2.1 in-silico validation pipeline: molecule generation, ADMET profiling, retrosynthesis check, safety prefilter, rigid and induced-fit docking, selectivity panel, MD, free-energy, orthogonal validation, and lab-in-the-loop handoff. The table below shows every hit — paginated — with a status dot per stage.

What this means for researchersThese are the compounds that passed virtual screening with positive DiffDock confidence scores(> 0), meaning the AI predicts they will physically bind to SMA-relevant protein targets. Hits are ranked by confidence score — higher is better. The multi-stage pipeline tracks each hit from molecule generation through docking, MD, and free-energy to wet-lab handoff. Green dots = completed stage, yellow = in progress, gray = pending, faint/outlined = not applicable.

Confidence score guide:> +0.5 = high-confidence binder (strong signal), +0.1 to +0.5 = moderate binder, 0 to +0.1 = marginal (needs validation). Scores below 0 are filtered out and not shown. All scores are computational predictions; no wet-lab validation has been performed.

Note: ADMET predictions in the pipeline use rule-based heuristics (Lipinski, TPSA, PAINS), not validated PK/tox models.

▶How does Hit Validation work?

Confidence score guide: >+0.5 = high-confidence binder, +0.1 to +0.5 = moderate binder, 0 to +0.1 = marginal. Scores below 0 are filtered out. All scores are computational predictions only; no wet-lab validation has been performed.

Stage dots: completed · in progress · pending.

ADMET predictions use rule-based heuristics (Lipinski, TPSA, PAINS), not validated PK/tox models.

Loading screening hits…