The Platform

Full simulation pipeline. One ranked output.

Three simulation layers working in concert: molecular docking pre-filter, alchemical FEP binding affinity ranking, and 12-property ADMET scoring — producing a single ranked synthesis queue your medicinal chemistry team can act on immediately.

Request Access See the Science

DrugSynq platform pipeline diagram showing molecule input flowing through FEP binding affinity scoring and ADMET prediction to ranked synthesis output

Architecture

Three layers. No black boxes.

Every step in the DrugSynq pipeline is grounded in well-established physics or validated ML. We don't apply methods you can't interrogate.

Layer 1 — Docking Pre-filter

High-throughput rigid docking narrows your library from 100K candidates to the top 500–1,000 by binding pose score. Computationally inexpensive; sets up the FEP input set.

Layer 2 — Alchemical FEP Ranking

Relative binding free energy calculations using OPLS4 force field and Hamiltonian replica exchange. ΔΔG values ranked across the candidate set with estimated statistical error.

Layer 3 — ADMET Scoring

12 ADMET properties predicted using ensemble ML models trained on in vitro assay data. Properties include permeability, hERG, CYP inhibition, metabolic stability, and aqueous solubility.

Binding Affinity Ranking

ΔG predictions with physics, not pattern-matching.

Free-energy perturbation calculates the thermodynamic work of alchemically transforming one ligand into another within the protein binding site. This thermodynamic rigor translates to better extrapolation into unexplored SAR space — critical for lead optimization where you're often synthesizing analogs with no prior activity data.

OPLS4 / AMBER force field with custom small-molecule parameters

Hamiltonian replica exchange enhanced sampling — avoids conformational trapping

Relative ΔΔG with estimated statistical uncertainty per compound

Supports congeneric and scaffold-hop perturbation networks

ADMET Engine

Twelve properties. One synthesis risk score.

ADMET failure is the leading cause of late-stage compound attrition. DrugSynq scores 12 properties in parallel with FEP ranking so you never optimize a compound into a metabolic liability or hERG blocker.

Caco-2 Permeability

hERG Inhibition

CYP3A4 Inhibition

CYP2D6 Inhibition

Metabolic Stability (HLM)

Aqueous Solubility (pH 7.4)

Plasma Protein Binding

BBB Penetration

P-gp Substrate

AMES Mutagenicity

LogP / LogD

Oral Bioavailability (F)

Sample ADMET Report

hERG Inhibition HIGH RISK

CYP3A4 Inhibition MODERATE

Metabolic Stability GOOD

Aqueous Solubility GOOD

Caco-2 Permeability HIGH

Molecule Prioritization

Multi-parameter optimization surface.

Binding affinity and ADMET properties are combined into a configurable multi-parameter optimization (MPO) score that weights properties by your program's specific requirements. Export as CSV, JSON, or push directly to your ELN.

Configurable MPO Weights

Adjust relative weights for binding affinity, individual ADMET properties, and synthetic accessibility to match program goals. CNS programs weight BBB differently from oncology.

Ranked Synthesis Queue

Output is a single ranked list sorted by composite MPO score, with per-compound breakdowns for each property flag. Decision-ready, not data-dump.

Multiple Output Formats

Download results as CSV with SMILES, predicted ΔG, all 12 ADMET properties, and MPO rank. JSON output for downstream computational workflows. ELN integration via API.

API Access

Integrate DrugSynq into your computational workflow.

REST API available on Discovery and Pipeline tiers. Authenticate with API key, submit job via POST, poll for results or receive webhook notification on completion.

DrugSynq REST API — Submit Ranking Job

# POST /v1/jobs — submit molecule ranking job
$ curl -X POST https://api.drugsynq.com/v1/jobs \
  -H "Authorization: Bearer dsq_api_key_..." \
  -H "Content-Type: application/json" \
  -d '{
    "target_pdb": "7L11",
    "molecules": [
      {"smiles": "Cc1ccc(NC(=O)c2ccc(CN3CCN(C)CC3)cc2)cc1Nc1nccc(-c2cccnc2)n1", "id": "mol_001"},
      {"smiles": "CC(C)(C)c1ccc2c(c1)CC(=O)N2c1cc(Cl)cc(Cl)c1", "id": "mol_002"}
    ],
    "fep_mode": "relative",
    "admet": true,
    "mpo_weights": {"affinity": 0.6, "admet_composite": 0.4}
  }'

# Response:
{
  "job_id": "dsq_job_8f2a91c",
  "status": "queued",
  "estimated_completion": "2026-06-09T16:30:00Z",
  "webhook_url": null
}

# GET /v1/jobs/{id}/results — retrieve ranked output
$ curl https://api.drugsynq.com/v1/jobs/dsq_job_8f2a91c/results \
  -H "Authorization: Bearer dsq_api_key_..."

{
  "job_id": "dsq_job_8f2a91c",
  "status": "complete",
  "results": [
    {"id": "mol_002", "delta_g": -9.34, "mpo_rank": 1, "admet_flags": [], "delta_g_err": 0.42},
    {"id": "mol_001", "delta_g": -8.71, "mpo_rank": 2, "admet_flags": ["herg_moderate"], "delta_g_err": 0.38}
  ]
}