Engineering the next generation of small-molecule medicines
Talia Research Center unites generative chemistry, machine-learning property prediction and computational genomics to design brain-penetrant therapeutics for diseases with no targeted treatment.
Research areas
Four programs, one computational core
We pair deep disease biology with AI-driven molecular design across discovery and translational science.
AI Drug Discovery & Molecular Design
Generative chemistry and message-passing neural networks (D-MPNN) design and rank novel candidate analogs, predicting potency and ADMET properties before a single compound is made.
Targeted Small-Molecule Therapeutics
Structure-based optimization of STING inhibitors and kinase programs (Nek, CLK2), advancing series with improved selectivity, potency and developability.
Computational Genomics & Precision Oncology
Whole-genome and tumor–normal paired analysis — SNP, InDel, SV and CNV calling with somatic variant interpretation — to nominate targets and stratify disease.
Neuroinflammation & CNS Biology
Brain-penetrant therapeutics for interferonopathies such as AGS, SAVI and neuropsychiatric lupus, with long-term opportunities in neurodegeneration.
The platform
An AI-enabled optimization engine
Every design cycle is closed-loop: models generate, rank and filter candidates, then learn from each round of experimental data.
Generative Chemistry & ML
AI-driven generation of candidate analogs and SMILES libraries with predicted target activity.
- Novel analog generation
- Structure–activity learning
- Predicted potency (pIC50)
ML Activity & Property Ranking
A ranked, review-ready shortlist scored on activity, CNS/ADME properties and synthetic feasibility.
- Ranked candidate shortlist
- CNS / ADME modeling
- Synthetic feasibility scoring
Multi-Parameter Optimization
Simultaneous optimization across the properties that make or break a CNS drug.
- BBB & toxicity modeling
- MDR1 / BCRP efflux filtering
- hERG & off-target risk
How we work
A closed-loop, active-learning cycle
Design–Make–Test–Analyze iterations compound model accuracy with every generation.
Design
Generative models propose novel analogs optimized against multiple objectives.
Synthesize
Top-ranked, synthetically feasible candidates are prioritized and made.
Test
In-cell target engagement, IC50, ADME and CNS-exposure assays measure real activity.
Analyze
Results retrain the models, sharpening the next round of predictions.
Model accuracy across generations
Active learning lifted predictive performance from benchmark to sub-micromolar candidates.
Pipeline
Programs in progress
A focused portfolio spanning neuroinflammation, oncology and precision-medicine platforms.
| Program | Indication | Modality | Stage |
|---|---|---|---|
| TRC-S1 — STING inhibitor series | Neuroinflammation (AGS, SAVI) | Brain-penetrant small molecule | Lead Opt |
| TRC-NP — NPSLE program | Neuropsychiatric lupus | Small molecule | Discovery |
| TRC-N1 — Nek kinase program | Oncology (lymphoma) | Small-molecule kinase inhibitor | Discovery |
| TRC-C1 — CLK2 modulator | Oncology | Small molecule | Discovery |
| TRC-G1 — Genomics platform | Precision oncology | WGS variant analytics | Research |
Pipeline shown for illustration of research scope; stages reflect internal discovery status.
About
A research center built around computation
Talia Research Center is an independent biomedical research organization advancing AI-native drug discovery. We bring machine learning, medicinal chemistry, structural biology and genomics under one roof to attack diseases that conventional approaches have left behind.
We collaborate with industry and academic partners — including Zermatt Biotech — to translate computational predictions into validated, clinic-ready candidates.
Disease-first
We start from unmet need and rare, genetically-defined disease, then engineer the molecule to fit.
Open collaboration
Shared platforms and partnerships accelerate the path from hypothesis to candidate.
Translational rigor
Every prediction is held to experimental and developability standards before it advances.
Let's advance the science together
Partner with us on target discovery, AI-driven optimization or translational collaboration.