The end-to-end agentic platform for drug discovery and development

Five connected research stages, every result traceable to its data source and computation. Observations, predictions and conditional hypotheses stay clearly separated, and the outcome is an interactive workspace with costed selective reruns rather than a static report.

1. 01

   Mechanism, literature & omics intelligence

   From a disease, phenotype, target or molecule to a ranked, evidence-backed target hypothesis — indication-independent, for every organ.

   • 1.1 Literature mining

     PubMed and open-access full-text review of top-impact journals with structured extraction of findings and traceable references.

   • 1.2 Omics analysis

     Transcriptomics cohorts from GEO, differential expression, disease signatures and human tissue context (Human Protein Atlas).

   • 1.3 Target identification & prioritisation

     Causal mechanism chains, Orphanet ontologies, genetics and druggability on one evidence board; first-in-class and drug-repurposing opportunities.

2. 02

   Therapeutic design

   Multimodal design of new chemical and biological entities against the selected targets.

   • 2.1 Protein structure prediction

     Target and complex structures with Boltz-2 and qualified experimental PDB structures.

   • 2.2 Small-molecule design

     Generative de novo design and REINVENT4 Mol2Mol optimisation grounded in ChEMBL actives.

   • 2.3 Peptide, protein & antibody design

     Binder design for peptides, proteins and antibodies with documented wet-lab binding references.

3. 03

   Preclinical computation

   GPU physics and in-silico pharmacology that explain what each result means in the wet lab.

   • 3.1 Molecular docking & complex prediction

     Binding poses, protein–ligand and protein–protein complexes.

   • 3.2 Molecular dynamics (OpenMM)

     Complex stability, binding free-energy estimates and conformational dynamics on GPU.

   • 3.3 ADMET & safety

     Absorption, distribution, metabolism, excretion and toxicity predictions, hERG and off-target safety panels.

   • 3.4 Blood–brain barrier & membrane permeation

     BBB models, membrane simulation and permeability for CNS and peripheral targets.

   • 3.5 Route of administration & dose

     Exposure-driven route and dose proposals for every carried candidate.

   • 3.6 Network biology & systems pharmacology

     STRING, OmniPath and LINCS downstream signalling, QSP simulation and the cell-level mechanism hypothesis.

4. 04

   Clinical-development scenarios

   Qualified, clearly labelled in-silico scenarios for the path into humans.

   • 4.1 Clinical scenario

     Indication, population, comparator and endpoint framing.

   • 4.2 Human PK & virtual patients

     Population pharmacokinetics on GPU with NHANES virtual adults; TMDD for antibodies and proteins.

   • 4.3 Drug–drug interactions & pharmacogenomics

     FDA CYP interaction models, real co-medications and CPIC CYP2D6/2C19/2C9/3A5 phenotypes.

   • 4.4 Whole-body PBPK & special populations

     In development

     Physiologically based PK for organ exposure, renal and hepatic impairment.

   • 4.5 Organ safety

     In development

     Organ-level safety assessment from exposure and pharmacology.

   • 4.6 First-in-human dose & therapeutic window

     In development

     MABEL/NOAEL-based FIH starting dose and exposure margins.

   • 4.7 Disease QSP, stratification & virtual trials

     In development

     Disease progression models, patient stratification, virtual clinical trials and dose escalation.

   • 4.8 Regulatory dossier

     In development

     Structured evidence package for regulatory discussion.

5. 05

   Translation & economic assessment

   Long-term outcomes, translational risk and the economic case of a programme.

   • 5.1 Translational risk

     In development

     Evidence gaps, validation experiments and go/no-go criteria.

   • 5.2 Health economics

     In development

     Development cost, market and value assessment.