METAWORK combines deep learning, molecular simulation, and generative chemistry to engineer small-molecule therapeutics for historically challenging biological targets, accelerating the path from concept to clinic for difficult-to-treat cancers.
The METAWORK platform represents a paradigm shift in computational drug discovery. By integrating proprietary generative AI models with advanced molecular dynamics simulations, we can predict and optimize drug-protein binding interactions with unprecedented accuracy, enabling the rapid design of small-molecule therapeutics against targets once considered undruggable.
Our proprietary foundation model predicts three-dimensional binding conformations of small molecules against protein targets with sub-angstrom resolution, outperforming traditional docking methods by orders of magnitude in both speed and accuracy.
Simultaneously optimize potency, selectivity, metabolic stability, membrane permeability, and other ADMET properties through our integrated machine learning framework, dramatically reducing late-stage attrition in drug development.
Our de novo molecular generation system creates novel chemical scaffolds optimized for specific biological targets, exploring vast regions of chemical space inaccessible to traditional medicinal chemistry approaches while maintaining synthetic feasibility.
Transforming the landscape of drug discovery through the convergence of artificial intelligence, computational chemistry, and deep biological understanding to unlock treatments for the most challenging diseases.
METAWORK's integrated computational platform combines three synergistic pillars of technology: deep learning for structure prediction, physics-based molecular simulation for validation, and generative AI for chemical design. This unique combination enables us to navigate the vast landscape of drug-like chemistry with unprecedented precision and efficiency.
Our proprietary foundation model, trained on over 500 million protein-ligand interactions, predicts binding poses and affinities with state-of-the-art accuracy. The model captures complex quantum mechanical effects and entropic contributions that traditional docking methods miss entirely, enabling accurate predictions even for novel chemical scaffolds and challenging binding sites with shallow, dynamic, or allosteric characteristics.
High-fidelity molecular dynamics simulations at microsecond timescales validate and refine AI-generated predictions through explicit modeling of protein flexibility, solvation effects, and binding kinetics. Our GPU-accelerated simulation infrastructure can evaluate hundreds of compounds per day, providing critical data on binding residence time and conformational dynamics that correlate strongly with in vivo efficacy.
Our reinforcement learning-based generative model designs novel molecules optimized for multiple objectives simultaneously, including target affinity, selectivity against off-targets, metabolic stability, solubility, and synthetic accessibility. The system generates compounds that are not only computationally optimized but also practically synthesizable, with over 85% of generated compounds successfully made in the laboratory on first attempt.
Comprehensive machine learning models predict absorption, distribution, metabolism, excretion, and toxicity properties with accuracy validated against extensive experimental datasets. Early integration of ADMET optimization into the design cycle dramatically reduces late-stage failures and ensures that generated compounds possess favorable drug-like properties suitable for human therapeutic use.
Our diversified pipeline targets multiple oncology indications with a focus on historically challenging targets. Each program leverages our AI-driven platform to identify and optimize novel small molecules with differentiated mechanisms of action and improved therapeutic profiles.
A selective covalent inhibitor targeting the KRAS G12D mutation, one of the most common oncogenic drivers in pancreatic, colorectal, and lung cancers. Our compound demonstrates sub-nanomolar potency with exceptional selectivity over wild-type KRAS, addressing a significant unmet medical need in precision oncology.
A first-in-class molecular glue degrader targeting WRN helicase, exploiting synthetic lethality in microsatellite instability-high tumors. This program represents a novel therapeutic approach for colorectal, endometrial, and gastric cancers with MSI-H phenotype, potentially benefiting over 150,000 patients annually.
A highly selective dual CDK12/13 inhibitor designed to induce transcriptional stress and activate innate immune responses in HR-proficient tumors. This mechanism creates synthetic lethality in BRCA wild-type ovarian and prostate cancers, expanding immunotherapy responsiveness.
A tissue-selective PROTAC targeting BCL-XL for degradation specifically within tumor cells, overcoming the on-target thrombocytopenia that has limited BCL-2 family inhibitors. Our platelet-sparing approach enables dose intensification and improved therapeutic window in solid tumors.
An allosteric inhibitor of USP7 deubiquitinase that stabilizes p53 and PTEN tumor suppressors without the toxicity associated with catalytic site inhibitors. Our AI platform identified a novel allosteric pocket enabling selective modulation of oncogenic substrates while sparing essential cellular functions.
A small molecule that disrupts the MYC-MAX protein-protein interaction, targeting the transcription factor considered undruggable for decades. Our compound stabilizes an alternative MYC conformation that prevents DNA binding, demonstrating efficacy across multiple MYC-driven tumor models including neuroblastoma and medulloblastoma.
From target selection to clinical candidate nomination, our integrated computational platform accelerates every stage of drug discovery while maintaining the rigor required for successful translation to human therapeutics.
Deep structural analysis and binding site identification using our AI models to understand target dynamics, cryptic pockets, and allosteric mechanisms that inform druggability assessment and strategy selection.
AI-driven exploration of chemical space to generate novel molecular scaffolds optimized for target engagement, selectivity, and drug-like properties, producing thousands of high-quality candidates in days rather than months.
Simultaneous optimization of potency, selectivity, ADMET properties, and synthetic feasibility through active learning cycles that integrate computational predictions with experimental feedback.
Rigorous validation of lead candidates through in vitro pharmacology, pharmacokinetic studies, and efficacy models to ensure robust translation from computational predictions to clinical success.
We concentrate our efforts on oncology indications where our AI platform can make the greatest impact: difficult-to-treat cancers driven by historically undruggable targets, resistance mechanisms that evade current therapies, and rare molecular subtypes lacking effective treatment options.
KRAS mutations occur in approximately 25% of all human cancers, with historically limited therapeutic options due to the shallow binding pocket and high GTP affinity. Our generative AI platform has enabled the discovery of novel covalent and non-covalent inhibitors against multiple KRAS mutant alleles, including G12D, G12V, and G13C, with selectivity profiles suitable for chronic dosing in patients with pancreatic ductal adenocarcinoma, non-small cell lung cancer, and colorectal carcinoma.
Synthetic lethality approaches exploiting defects in DNA repair pathways offer transformative opportunities in oncology. Our platform identifies novel inhibitors of DNA damage response proteins including WRN helicase, POLQ, and RAD52 that create synthetic lethal relationships in tumors with specific genomic backgrounds, extending the PARP inhibitor paradigm to new patient populations with HR-proficient tumors.
Transcription factors drive tumorigenesis in many cancers but have been considered undruggable due to their lack of enzymatic activity and reliance on protein-protein interactions. Our AI platform has enabled breakthrough discoveries in targeting transcription factor complexes including MYC-MAX, STAT3, and AR splice variants through novel mechanisms that stabilize inactive conformations or disrupt essential co-factor interactions.
PROTAC and molecular glue degraders offer the ability to eliminate disease-causing proteins entirely rather than merely inhibiting their function. Our platform designs tissue-selective degraders and predicts ternary complex formation to enable degradation of targets that cannot be addressed by traditional small molecule inhibitors, including scaffolding proteins, transcription factors, and proteins lacking functional catalytic domains.
METAWORK combines world-class computational capabilities with deep drug discovery expertise to deliver differentiated therapeutics against the most challenging targets in oncology.
Proprietary Foundation Model trained on the largest curated dataset of protein-ligand structures, achieving sub-angstrom accuracy in binding pose prediction across diverse target families.
End-to-End Integration from target validation through clinical candidate selection, with seamless handoff between computational design, medicinal chemistry, and biological validation teams.
Validated Track Record with multiple programs advancing through preclinical development and partnerships with three major pharmaceutical companies validating our platform capabilities.
Multi-Modal Optimization simultaneously balancing potency, selectivity, ADMET properties, and synthetic accessibility through our unified machine learning framework.
Undruggable Target Expertise with specialized capabilities for protein-protein interactions, allosteric sites, dynamic binding pockets, and intrinsically disordered regions.
Rapid Design Cycles generating and evaluating millions of compounds in silico before any synthesis, reducing time from target to clinical candidate by 50% compared to traditional approaches.
Whether you're interested in partnership opportunities, exploring how our AI platform can accelerate your drug discovery programs, or learning more about our pipeline, we'd love to hear from you. Our business development team is ready to discuss how METAWORK can help you bring transformative therapeutics to patients faster.
1233 South Westgate Avenue
Apt 414
Los Angeles, CA 90025