Artificial intelligence to expand the notion of drug design

Primary in silico and in vitro target-specific selectivity assessment.

Hit Identification

We utilize the most efficient hit identification methodology combining AI techniques, traditional in silico approaches and feedback from experimental evaluation, which provides favorable candidates for further lead selection and optimization.

Primary in silico and in vitro toxicity assessment.

Target molecule preparation (homology and ab initio modeling, structural quality assessment, binding sites identification).

Generation of ultra-large chemical databases with specified properties.

Ultra-fast chemical space exploration (similarity searching, shape and electrostatic alignment and scoring).

Curated custom and focused chemical libraries (DNA-encoded, fragments, diversity, natural products, etc.).

Small molecules preparation and augmentation.

Screening of large-scale multi billion chemical databases.

Drug-target interaction prediction using graph neural networks.

AI-assisted ligand- and structure-based virtual screening.

Rapid synthesis of designed hit compounds.

Molecular docking with custom-tuned AI scoring functions.

Assays for hit evaluation, in vitro screening and profiling.

Safety, affinity and selectivity of the top-tier lead compounds are maximized by adjusting their chemical structures with AI-based drug optimization methods, synthetic chemistry and iterative experimental assessment. This efficiently turns the leads into drug candidates ready for preclinical and clinical testing.

Lead Optimization

Comprehensive selectivity profiling with selected organ-specific targets.

Automated lead annotation.

The best hit compounds are expanded into the series of derivative compounds which are assessed for pharmacokinetic and toxicological parameters with multiple AI models tuned by experimental feedback. Our approach eliminates dead-end structures and reduces time and costs of lead development significantly.

Functional cellular assays.

Experimental biochemistry and enzymology.

Hit-to-Lead

Exploration of the structure-activity relationships (SAR).

Experimental feedback to improve bioactivity.

Off-target interactions evaluation using drug-target affinity predictive AI models.

Structure-based optimization (using AI and SAR information).

Computer-aided drug design (in silico hit derivatives expansion and diversification, rapid synthesis and in vitro tests).

ADME-Tox prediction (toxicity class identification, cytotoxicity, hepatotoxicity, acute toxicity, carcinogenicity, plasma protein binding, oral bioavailability, cytochromes inhibition, etc.).

Advanced in silico and in vitro target-specific selectivity assessment.

Molecular dynamics and quantum calculations.

Optimization of the structure-activity and structure-toxicity relationships.

Experimental feedback to improve safety.

Advanced medicinal chemistry and molecular optimization.

Comprehensive polypharmacological profiles of the leads via knowledge graph.

Solubility and permeability fine-tuning to increase bioavailability

Validation by molecular dynamics simulations and quantum chemistry techniques.

Advanced ADME-Tox prediction (metabolism, biotransformation, biodistribution, compound excretion, Ames mutagenicity, etc.).

Ames test, Irwin's test, high-dose pharmacology, repeated dose toxicity in animal models, pharmacokinetics/ pharmacodynamics (PK/PD) studies, drug-induced metabolism exploration.

Confirmation of on-target and off-target effects.

Drug Candidate

Safety

Efficacy

Affinity

Selectivity

Stability

Bioavailability

A flexible modularity of our AI services can be configured to solve challenging tasks
via building up a drug discovery pipeline of any complexity.

Artificial Intelligence for Drug Design