Screening platform

The Moleculepath Screening Platform

Three integrated capabilities: structural screening, fragment library design, and hit triage — all anchored to the target protein structure your team already has.

Overview

Structure-first by design, not by default.

Most virtual screening platforms treat the protein structure as one input among many. Moleculepath is built around a different premise: the binding site geometry is the constraint that determines which compounds are worth evaluating. Every parameter — fragment set composition, docking protocol, scoring function, ADMET filter threshold — is set after pocket analysis, not before.

The platform accepts: a target protein structure (crystal, cryo-EM, or homology model with confidence metrics), any known ligand data or co-crystal structures, and program constraints (selectivity requirements, ADMET flags, IP freedom-to-operate notes). From that input, Moleculepath returns a curated hit shortlist with binding pose rationalization.

Module 01

Structural Screening

Binding site analysis maps the geometry, electrostatics, and flexibility of the pocket before a single compound is evaluated. The DFG loop conformation, hinge geometry, gatekeeper residue, and allosteric pocket availability are characterized from the input structure. Docking protocols are then calibrated to the specific pocket architecture — scoring functions, sampling parameters, and pose-filter criteria are all pocket-specific.

Docking is performed using validated physics-based methods (AutoDock Vina, GOLD, Glide-equivalent scoring). Pose clusters are analyzed for binding mode consistency before shortlisting. Not every top-scored compound survives pose analysis — structural consistency is a filter, not an afterthought.

Module 02

Fragment Library Design

Fragment sets are not selected from a catalog. They are computationally assembled for each target by filtering a curated fragment pool for pocket-compatible physicochemistry: molecular weight ≤300 Da, cLogP ≤3, and at least one specific pharmacophoric feature matching a key contact in the binding site.

For kinase targets, this means hinge-binding pharmacophores with defined vectors into the hydrophobic back pocket. For GPCR orthosteric sites, it means lipophilic fragments compatible with the extracellular loop geometry. Library size is typically 500–2,000 fragments per target, far smaller than a diversity screen — which is the point.

Module 03

Hit Triage Engine

The triage output is a ranked shortlist — not a CSV of docking scores. Each compound in the shortlist is accompanied by: a 3D binding pose visualization showing key interactions; a docking score distribution showing where this compound falls relative to the full screened set; ADMET profile (solubility prediction, metabolic liability flags, CYP inhibition risk, permeability estimate); and a written structural rationale explaining why this compound was retained at triage.

For top-ranked hits, FEP perturbation is applied selectively when structural confidence warrants — not as a default on all compounds. Delivery format: PDF partner data package + SDF file + optional interactive visualization session.

Data delivery

What you receive at the end of a campaign.

Every Moleculepath partner data package includes the following deliverables, formatted for direct use by the medicinal chemistry team:

  • 01 Ranked hit shortlist PDF — 15–30 compounds with binding pose 3D visualization, docking score, ADMET summary, and a written structural rationale per compound explaining the selection basis.
  • 02 SDF file — 3D coordinates for all shortlisted compounds, ready for import into Schrödinger, MOE, or internal cheminformatics pipelines. Includes docking score and key contact annotations in SDF properties.
  • 03 Campaign design summary — binding site analysis results, fragment library selection rationale, docking protocol parameters, scoring function calibration on any available reference ligands.
  • 04 Interactive visualization session — 60-minute video call to walk through binding poses, answer structural questions, and discuss hit-to-lead prioritization with the Moleculepath computational chemistry team.

Partner data is handled under mutual NDA. Structural data and screening results are treated as proprietary program information and are not shared outside the partnership.

Ready to put your target structure to work?

A target briefing takes 30 minutes. From that conversation, Moleculepath returns a campaign design proposal within five business days.

Discuss a Target