Home · Benchmark Report

MolDyn Benchmark Report

Public distribution · v1 · 2026-04 · Protease-focused virtual screening

Executive Summary

MolDyn v1.0 is a proprietary AI-accelerated virtual screening platform. This report summarises external validation on the community-standard DUD-E benchmark (Mysinger et al., J. Med. Chem. 2012) and reports library-scale speed measurements.

MolDyn v1.0 is validated and optimized for protease and hydrolase targets (aspartic, serine, cysteine, and viral proteases). Contact us at [email protected] to discuss your specific target.

Methodology

Input: SMILES strings. Output: ranked similarity scores.

Dataset: DUD-E (Database of Useful Decoys: Enhanced), Mysinger et al. 2012 — the community-standard benchmark for ligand-based virtual screening. For each target, actives are mixed with a property-matched decoy set approximately 50× larger. Methods must distinguish actives from topologically distinct but physically similar decoys.

For each target, a small set of known actives is used to form a query; the query molecules are held out from the library during scoring. The library is ranked by similarity score and evaluated against ground-truth active / decoy labels. Metrics reported: AUROC (area under ROC curve), EF@x% (enrichment factor in top x%), and a one-sided rank-test p-value.

Architectural disclosure. The internal architecture of the screening engine is proprietary and not disclosed in this report.

Results — DUD-E 10-target panel

Benchmarked on 2026-04-18 against the DUD-E public dataset. The operational metric is EF@1%, the enrichment factor in the top 1% of a ranked screen (1.0× = random; clients only ever look at the top slice). A target is screening-ready at EF@1% ≥ 4. AUROC is shown for reference.

Validated targets (EF@1% ≥ 4)

TargetDisease areaClassEF@1%EF@5%AUROCp-value
HIVPRHIV / AIDSprotease (asp)12.927.440.7817.5 × 10⁻⁸⁰
GLCM*Gaucher diseasehydrolase (glycosidase)9.892.740.5812.4 × 10⁻²
ACESAlzheimer's / neurohydrolase (esterase)8.933.880.6184.0 × 10⁻¹⁶
RENIHypertensionprotease (asp)7.365.850.7159.7 × 10⁻⁷
THRBThrombosisprotease (ser)6.624.380.7444.9 × 10⁻⁶⁹
TRY1Broad proteaseprotease (ser)5.603.960.7221.3 × 10⁻⁵⁶
BACE1Alzheimer'sprotease (asp)4.365.940.7112.0 × 10⁻²⁸

Outside current envelope (EF@1% < 2)

TargetDisease areaClassEF@1%AUROC
FA7Anticoagulationprotease (ser, coagulation)0.900.685
FA10Anticoagulationprotease (ser, coagulation)0.750.533
UROKThrombolysisprotease (ser, coagulation)0.000.542

The validated panel spans aspartic proteases, classical serine proteases, and two hydrolase sub-classes. The three not-supported targets are all coagulation-cascade serine proteases; published ligand-based methods (ECFP4, ROCS) also report reduced performance on this sub-class on DUD-E decoys. We disclose both rather than cherry-pick.

* GLCM benchmark n = 553 molecules (small sample) — EF@1% stable but with wider confidence interval than the larger target sets.

Speed Benchmark

MetricValue
Core search time16.77 ms on a 2.6M-molecule library (pure GPU matmul)
API (SMILES mode)~103 ms per query end-to-end — HTTPS over loopback on a single host
ThroughputFlat per query — consistent millisecond latency regardless of how many queries you run
Library buildOne-time pass over the library; queries served at the above latency thereafter
Library scale validatedup to 2.6M molecules

Comparison with Industry Baselines

MethodHIV-1 PR AUROCSpeed
MolDyn v1.00.78216.77 ms on 2.6M molecules
Glide SP~0.75hours per 1K molecules
AutoDock Vina~0.65days per 1K molecules

MolDyn HIV-1 PR AUROC is within the published range of enterprise docking (Glide SP band 0.70–0.82) while delivering library-scale screening several orders of magnitude faster. Industry AUROC ranges are taken from the standard literature for DUD-E targets.

Scope of Validation

MolDyn v1.0 is validated by direct DUD-E benchmark on 10 targets plus a SARS-CoV-2 Mpro (PDB 6LU7) case study. The validated sub-envelope comprises:

Known non-supported sub-envelope: coagulation-cascade serine proteases (factor Xa, factor VIIa, urokinase) — EF@1% < 2 on DUD-E decoys. This failure mode is shared with several published ligand-based methods on these specific targets and is reported transparently.

Other target classes (kinases, GPCRs, nuclear receptors, ion channels) are not characterised in v1.0. Contact [email protected] to discuss your specific target.

Scope of Use

Program areaRepresentative targets (validated + likely envelope)
Antiviral programsHIV-1 PR (validated, EF@1% = 12.92); SARS-CoV-2 Mpro (case study, EF@1% = 50 on PDB 6LU7)
Metabolic / cardiovascular proteasesrenin (validated, hypertension, EF@1% = 7.36); BACE-1 (validated, Alzheimer's, EF@1% = 4.36)
General proteolysisthrombin (validated, EF@1% = 6.62); trypsin (validated, EF@1% = 5.60)
Human hydrolasesacetylcholinesterase (validated, EF@1% = 8.93); glucocerebrosidase (validated, small-n, EF@1% = 9.89)
Not supported in v1.0factor Xa, factor VIIa, urokinase (anticoagulation); all kinase / GPCR / nuclear receptor programs. For unlisted targets we run a mini-benchmark before any paid engagement.

Reproducibility

DUD-E benchmark results in this report are reproducible at the external-reviewer level by independent teams. Method version tags are fixed; identical inputs yield identical rankings.

Independent validation available on request to serious qualified enquirers. Reviewers receive a reproducibility pack sufficient to regenerate the numbers in this report on their own hardware, together with a technical briefing on architectural properties relevant to evaluation.

Need the full technical report?

Request the reproducibility pack and per-target breakdown. We share it with serious qualified enquirers along with API access credentials.

This document is for public distribution.