Who should use the Design new molecules for drug candidates workflow?
Teams or solo builders working on science & healthcare tasks who want a repeatable process instead of one-off tool experiments.
AI Workflow · Science & Healthcare
Practical execution plan for design new molecules for drug candidates with clear steps, mapped tools, and delivery-focused outcomes.
Deliverable outcome
A final deliverable of 2–3 novel drug candidates with synthesis routes and full computational validation ready for experimental testing.
30-90 minutes
Includes setup plus initial result generation
Free to start
You can swap tools by pricing and policy requirements
A final deliverable of 2–3 novel drug candidates with synthesis routes and full computational validation ready for experimental testing.
Use each step output as the input for the next stage
Step map
Instead of relying on a single generic AI model, this pipeline connects specialized tools to maximize quality. First, you'll use AlphaFold to a clear target protein with a defined binding pocket ready for virtual screening or de novo design. Then, you pass the output to BenevolentAI (The Benevolent Platform™) to a virtual library of novel, synthetically plausible candidate molecules targeting the binding site. Then, you pass the output to Entos to a shortlist of 50–200 high-scoring molecules with predicted binding poses and interaction profiles. Then, you pass the output to Ersilia Open Source Initiative to a refined set of 20–50 drug-like molecules with predicted safety and pk profiles. Then, you pass the output to Entos to 3–5 optimized lead molecules with improved potency, selectivity, and drug-like properties. Then, you pass the output to Entos to confirmation of stable binding and dynamic behavior for 2–3 final lead molecules. Finally, Molecule.one is used to a final deliverable of 2–3 novel drug candidates with synthesis routes and full computational validation ready for experimental testing.
Define target protein and binding site
A clear target protein with a defined binding pocket ready for virtual screening or de novo design.
Generate initial molecular library or scaffold
A virtual library of novel, synthetically plausible candidate molecules targeting the binding site.
Perform virtual screening and docking
A shortlist of 50–200 high-scoring molecules with predicted binding poses and interaction profiles.
Evaluate ADMET and drug-likeness
A refined set of 20–50 drug-like molecules with predicted safety and PK profiles.
Optimize leads through iterative design
3–5 optimized lead molecules with improved potency, selectivity, and drug-like properties.
Validate with molecular dynamics simulations
Confirmation of stable binding and dynamic behavior for 2–3 final lead molecules.
Deliver final candidate list with synthesis plan
A final deliverable of 2–3 novel drug candidates with synthesis routes and full computational validation ready for experimental testing.
Identify the disease-relevant protein target and obtain its 3D structure (e.g., from PDB or AlphaFold). Map the binding pocket residues and key pharmacophoric features (hydrogen bond donors/acceptors, hydrophobic regions) to guide molecule design.
Why AlphaFold: AlphaFold provides protein structure prediction and drug target identification, directly supporting the definition of target protein and binding site.
Use generative AI models (e.g., REINVENT, MolGAN, or diffusion models) or fragment-based enumeration to create a diverse set of novel molecules that fit the binding pocket. Optionally, include known active scaffolds as seeds to bias generation.
Why BenevolentAI (The Benevolent Platform™): BenevolentAI offers molecular design and optimization, suitable for generating an initial molecular library or scaffold.
Dock the generated library into the target binding site using molecular docking software (e.g., AutoDock Vina, Glide). Score each molecule for binding affinity and pose quality, then select top 1–5% for further analysis.
Why Entos: Entos provides hit-to-lead screening and molecular property prediction, which can support virtual screening and docking workflows.
Compute absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties using QSAR models or rule-based filters (e.g., Lipinski’s Rule of Five, Veber rules). Prioritize molecules with favorable pharmacokinetic profiles and low toxicity risk.
Why Ersilia Open Source Initiative: Ersilia Open Source Initiative screens compound libraries for toxicity and ADME properties, directly addressing ADMET and drug-likeness evaluation.
Select top 5–10 molecules for iterative optimization: modify functional groups to improve potency, selectivity, or ADMET while maintaining synthetic feasibility. Use free-energy perturbation (FEP) or SAR analysis to guide changes.
Why Entos: Entos offers lead optimization and molecular property prediction, which aligns with iterative lead optimization.
Run explicit-solvent molecular dynamics (MD) simulations (100–500 ns) for top leads to assess binding stability, conformational changes, and off-target risks. Analyze RMSD, interaction persistence, and free energy landscapes.
Why Entos: Entos provides molecular property prediction and lead optimization, which can complement molecular dynamics validation.
Compile the top 2–3 molecules with detailed synthesis routes (retrosynthetic analysis using tools like Chematica or Reaxys), predicted spectra, and a summary of all computational data. Provide a report for experimental synthesis and testing.
Why Molecule.one: Molecule.one specializes in retrosynthesis planning and reaction condition recommendation, directly supporting delivery of a synthesis plan.
§ Before you start
Teams or solo builders working on science & healthcare tasks who want a repeatable process instead of one-off tool experiments.
No. Start with the top pick for each step, then replace tools only if they do not fit your pricing, compliance, or output needs.
Open the mapped task page and compare top options side by side. Prioritize output quality, integration fit, and predictable cost before scaling.
§ Related
Convert long-form videos into high-engagement short clips for TikTok, Reels, and YouTube Shorts automatically.
Launch a complete professional brand identity including logos, social assets, and marketing visuals using high-fidelity AI.
A complete end-to-end AI pipeline for generating video scripts, human-sounding voiceovers, and visual content — no camera or studio required.