Octara Drug
In developmentA computational molecular research workspace.
Evidence retrieval, molecular design and property modelling in one place. Octara picks the right engine for the modality and keeps every prediction attached to the model that produced it.
Read this first
Not an autonomous drug inventor
Octara Drug is a computational research environment that assists researchers with evidence retrieval, hypothesis generation, modelling and design. It does not discover drugs on its own.
Every molecule it produces is a virtual, computationally-generated ligand candidate. Every prediction is computational and requires experimental validation. Each panel shows its model, confidence and applicability domain, so a result can be judged rather than trusted.
The workflow
From a research question to an auditable record
A broad research workflow across modalities. The language model routes and explains; the engines and databases do the work.
- 01
Research question
A therapeutic goal or target, stated in plain language.
- 02
Evidence
Retrieval across target, structure and chemistry databases.
- 03
Target context
Druggability, essentiality and expression brought together.
- 04
Candidate design
Generative and library-based molecular design.
- 05
Molecular analysis
Physicochemistry, drug-likeness and novelty.
- 06
Property models
Docking, ADMET and selectivity, each with its model.
- 07
Medicinal chemistry
Iteration, selectivity reasoning and route planning.
- 08
Scientific record
An auditable result with confidence and provenance.
Modalities
One workspace, several kinds of molecule
Small molecule
de-novo, ADMET, docking, medchem
PROTAC degrader
E3 ligase, linker, ternary complex
Antibody
humanization, developability
Oligonucleotide
siRNA, ASO, off-target screen
Protein binder
de-novo miniprotein binder
Target discovery
druggability, essentiality, expression
Real output
Rendered from the application
Structures, routes and property panels produced by the engines, not illustrations.
