Mutation-aware docking

Find compounds that
prefer the mutant.

Pick a clinically relevant mutation, pick your compounds, and Liganx docks them against wild-type and the mutant in parallel — then shows you exactly which compounds gain selectivity. No PyMOL, no FoldX setup, no AutoDock wrangling.

No install Free Vina docking · Virtual screening on Pro Mutation-aware virtual screening Free for academic use
Sample
EGFR · selectivity matrix
example
CompoundWTT790MC797S
Osimertinib-8.8-9.3-5.2
Gefitinib-7.2-5.1-4.9
Erlotinib-7.6-5.4-5.3
Compound X-6.1-7.8-7.1
Compound X is predicted to bind the T790M mutant 1.7 kcal/mol better than wild-type — a candidate for resistance.
Vina scoring noise is roughly ±1 kcal/mol at default exhaustiveness — Δs above ~1 kcal/mol are interpretable, smaller deltas live near the noise floor.

Built on tools the community already trusts

AutoDock VinaRDKitPDBFixerMol*ProLIFRunPod
How it works

From mutation to selectivity matrix in three clicks.

1

Pick a target

Choose from clinically actionable kinases or upload your own PDB. Pocket boxes are pre-defined.

2

Pick mutations

Click EGFR T790M, KRAS G12C, BRAF V600E — anything from the curated library, or type your own.

3

Read the matrix

We dock every compound against WT and each mutant. The Δ-score shows you which compounds prefer which.

What's new

Recently shipped.

Just shipped

Resistance Atlas — predict the next mutation that breaks a drug

For every FDA-approved targeted cancer drug, the Atlas ranks which mutations will most likely emerge as clinical resistance — before patients hit them. Triangulates docking Δ + ESM-2 protein-language-model fitness, calibrated on 50 published clinical-resistance events (ROC-AUC 0.90 in-sample, 0.81 cross-validated). 15 drugs covered today. Novel (gene, position, mutant) lookups now run real ESM-2 on our GPU pod on demand.

Open the Atlas
Pro beta

Calibrate your own (drug, mutation) data against our model

Upload up to 10 (gene, position, wt, mutant, drug) rows as CSV. We score each through the same 2-signal model the Atlas uses — real ESM-2 inference for novel mutations, instant cache hits for the 49 calibration events — and return joint probability, verdict, and AUC if you provide ground truth. Free tier: 10 rows / day.

Calibrate your data
Just shipped

Pre-computed FDA-drug screenings — no setup, no GPU wait

We pre-ran 30 oncology kinase inhibitors against every resistance mutation in our catalog — KRAS G12C/G12D/Q61H, EGFR T790M/L858R/C797S, BCR-ABL T315I/E255K. Hit a public URL and see ranked selectivity hits in 1 second. Click any compound to see its 3D pose.

Browse pre-computed screenings
Just shipped

Mutation-aware virtual screening

Submit 10 compounds against a (target, mutation) pair. We dock each against WT and the mutant in parallel, rank by selectivity index, and return a ranked hit list. Real Vina, real poses, real ADMET. Sieve through a library; pick the top hits.

Open Studio
What you get

Everything an early-discovery med-chemist actually wants.

Selectivity matrix

N compounds × M mutants in one view, cells colored by Δ-score so resistance and selectivity gain pop out instantly. The whole product on one screen.

Pre-computed FDA-drug screenings

30 FDA-launched oncology kinase inhibitors pre-docked against every catalog resistance mutation — public landing pages, no login. Open any card on /library and see the ranked selectivity hits in one second.

Mutation-aware virtual screening

Drop up to 1000 compounds against a (target, mutation) pair; we dock each against WT and the mutant in parallel and return a hit list ranked by selectivity index. Promote the top hits to a full job in one click — no re-dock.

Resistance Atlas + calibrate your own data

A public per-drug atlas ranking the mutations most likely to break each FDA-approved targeted drug, calibrated on 50 clinical-resistance events (ROC-AUC 0.81 cross-validated). Upload your own (drug, mutation) CSV to score against the same ESM-2-backed model.

Validated poses, not just scores

Every job runs QuickVina2, with every pose checked by PoseBusters, Vinardo re-score, and RDKit strain analysis — most free tools give you no validation at all.

Honest by default

Every Δ near the ±1 kcal/mol Vina noise floor gets a within-noise badge; mutations outside the pocket are flagged, not scored; every pose comes with a plain-English readout and an inline ADMET panel (hERG, DILI, CYP, BBB). We publish our method limits on purpose.

Where we sit

The missing middle.

Free serversLiganxSchrödinger Maestro
Mutation-aware WT-vs-mutant matrixpartial
Public resistance-mutation atlas
Pre-computed FDA-drug screenings
Bulk virtual screening, selectivity-rankedpartial
Inline ADMET (hERG / DILI / CYP / BBB)partial
Ensemble / flexible-receptor dockingcoming soon
Runs in the browser, no install
Published, reproducible validation report

Reflects publicly known features as of May 2026. Free-server and Schrödinger capabilities vary by version, license tier, and module.

Method limitations we publish on purpose
See full validation report →
  • Eleven literature-anchored controls, public verdict. ABL T315I, EGFR T790M, BRAF V600E, KIT D816V, BTK C481S, KRAS G12C, EGFR C797S, EGFR L858R — five of eleven PASS in the published direction (four at above-noise magnitude, plus one correctly-retained within-noise case). Five NOISE results sit in documented method-limit territory (covalent acrylamides, active-conformation selectivity, conformational activation). One FAIL (EGFR L858R + Gefitinib) is explained candidly with the structural reason — rigid-receptor docking can't capture L858R's conformational activation. The full per-case verdict and the open-source script that re-derives it are public.
  • Vina noise floor. Vina/QuickVina2 scoring has roughly ±1 kcal/mol noise at default exhaustiveness. We surface a "within-noise" badge for any Δ inside that band so a reader doesn't over-interpret 0.3 kcal/mol shifts.
  • Mutant-receptor build path. Our default mutant builder is FoldX BuildModel where an academic licence permits it; on environments without FoldX we fall back to PDBFixer's residue substitution, which applies the new identity but does not energy-minimise the structure. Drastic side-chain changes (e.g. small→large) can introduce clash signal in the Δ that isn't pure binding affinity. Submitting the same mutation in both modes and comparing flags this when it matters. On liganx.com today FoldX is not deployed, so every public mutant is built with the PDBFixer path described here — the FoldX build path is available in the code for self-hosted or licensed deployments.
  • FoldX academic licensing. FoldX is free for academic use under its own EULA but requires a commercial licence for industry workflows. Liganx ships the FoldX call path; users running commercial work should verify their licence with the FoldX team at the Centre for Genomic Regulation directly.
  • Single-conformation rigid-receptor docking. We dock against one PDB conformation per (target, mutation). Mutations far from the binding pocket — typical for activation-loop, allosteric, or distant-domain residues — get an "outside pocket" badge instead of a possibly-misleading Δ, and the matrix shows them as not-scored rather than zero.
FAQ

Molecular docking, answered

What is molecular docking?+

Molecular docking predicts how a small molecule binds to a protein target and estimates the binding affinity (in kcal/mol). Liganx runs docking online in your browser using GPU-accelerated AutoDock Vina (QuickVina2) and the Boltz-2 machine-learning model, with no software to install.

Is Liganx molecular docking free?+

Yes. Liganx gives you free online molecular docking runs with no install and no credit card. You dock wild-type and mutant targets, view 3D poses, and export results. Paid tiers add bulk virtual screening of hundreds of compounds.

Do I need to install any software?+

No, Liganx is fully web-based. Pick a protein target (curated catalog, RCSB PDB search, or your own upload) and compounds (sketch, paste SMILES, or upload a CSV/SDF), and everything runs on cloud GPUs. Nothing to download or configure.

Which docking engine does Liganx use?+

Liganx docks with AutoDock Vina (QuickVina2-GPU), re-scores with Vinardo for sharper close-analog ranking, and also supports the Boltz-2 co-folding model. Every pose is checked with PoseBusters physics validation so you can trust the geometry.

What is mutation-aware docking?+

Mutation-aware docking compares how a compound binds the wild-type protein versus clinically relevant mutants, such as EGFR T790M, BCR-ABL T315I, BRAF V600E, or KRAS G12C, in a single run, so you can see which compounds keep or gain selectivity against drug-resistance mutations.

What do I need to run a docking?+

A protein structure (a 4-character PDB ID like 4OBE, an RCSB search, or an uploaded PDB) and one or more compounds as SMILES, a 2D sketch, or an uploaded file. Liganx handles receptor preparation, mutation building, docking, and pose scoring automatically.

Stop hand-rolling mutation-aware docking.

One UI. Real Vina under the hood. Selectivity matrix in minutes, not days.