HelixForge
HelixForgeby Bajpai Labs
Services

Three computational pipelines. Built for pharma timelines.

Every engagement returns a ranked candidate list plus the technical documentation your team needs to take work straight into the wet lab.

Service 01

Small Molecule Drug Discovery

Finding selective inhibitors for a disease target across millions of candidates.

Timeline
2–3 weeks
Starting price
Engagements typically $150K–$500K per program
Process
  1. 01Protein structure analysis (PDB or AlphaFold)
  2. 02Generate millions of candidate molecules with Graph Neural Networks
  3. 03Virtual docking + scoring simulations (AutoDock Vina, custom MD)
  4. 04Deliver top 50–100 molecules ranked by predicted efficacy
Deliverable
CSV with SMILES, binding affinity predictions, ADME properties
Service 02

Gene Therapy & Sequence Optimization

Designing DNA/RNA sequences for therapeutic delivery with high expression and low off-target risk.

Timeline
2–4 weeks
Starting price
Engagements typically $200K–$500K per program
Process
  1. 01Generate sequence variants with codon optimization & stability scoring
  2. 02Predict expression levels and off-target effects
  3. 03Assess manufacturability constraints
  4. 04Rank candidates across multiple objectives
Deliverable
Top 20 sequences + mutation map + expression predictions
Service 03

Antibody & Protein Engineering

Discovering antibodies against viral or cancer targets with strong developability.

Timeline
3–4 weeks
Starting price
Engagements typically $250K–$600K per program
Process
  1. 01Generate antibody sequence library (millions of variants)
  2. 02Predict binding affinity, thermostability, manufacturability
  3. 03Filter for developability metrics (aggregation, immunogenicity)
  4. 04Deliver top 100–500 antibodies ranked by predicted performance
Deliverable
Top 100–500 antibodies ranked by predicted performance
Methodology

The science behind the pipeline.

We compose published, peer-reviewed methods with proprietary scoring trained on our own wet-lab validation data.

Molecular docking

AutoDock Vina, DiffDock, and proprietary scoring functions.

ML models

Trained on ChEMBL, PubChem, BindingDB, and curated proprietary sets.

Molecular dynamics

GROMACS / OpenMM simulations for binding free-energy refinement.

Protein language models

ESM-2, ESMFold, and AlphaFold integration for structure & function.

Custom & Enterprise Projects

Multi-target optimization, integration with internal data lakes, real-time collaboration with your scientists, and bespoke ML model training on proprietary corpora.

  • Multi-target optimization
  • Integration with client data
  • Real-time scientist collaboration
  • Custom ML model training
Contact sales