AI Star - Pioneering the AI Discovery of Materials and Pharmaceuticals

Products

Swift Matrix

Deep Discovery

Drug Discovery AI Platform

Deep Discovery is a comprehensive web solution that provides reports on the three-dimensional binding structure between drugs and proteins by predicting the structure of proteins using protein sequence data and inference based on artificial intelligence.

Deep Discovery

Solution Modules

A solution that allows for the selection of drug candidate combinations by automatically collecting research big data and inferring using trained AI models.

Data Factory

Data Processing Module

Data Collection·Loading·Conversion
Feature Extraction
Data Preprocessing

Property Predictor

Feature Inference Module

Activation Energy
Interfacial Energy
Young's Modulus

Fastscan

New Drug Candidate Material Search Module

Candidate Proposal
Candidate Recommendation

Molender

Visualization Module

3D Protein Structure Visualization
Chart Visualization for Analysis

Unlike the conventional screening method, the solution provided by our company uses trained deep learning models to perform automated protein structure search and prediction.
Through this, we analyze the physical and chemical properties of the combined structure and candidate proteins to perform rapid and accurate candidate search.

Conventional Method

Drug candidates found via trial and error on limited data

DNA, RNA Sequence Analysis

Cryo-EM

X-ray Crystallization Analysis

NMR

Yeast Protein Synthesis

Researcher's guess

Months to years

High cost

When using our solution

Target Protein Prediction

Docking site

Optimization via physicochemical, bioinformatics, and molecular dynamics approaches
Structure prediction of drug candidates using training data

Hours needed

Low cost

Time
Workload
Costs

▼
▼
▼

Accuracy
Feasibility

▲
▲

Protein Folding

Amino Acid Sequence

Protein Structure

References

Bioinformatics

etc.

AI Training

Furthermore,
Using our own developed 6 core deep learning models, we provide accurate prediction of binding sites for drug targets.

Drug Database

O=C(NC!CCNCC1)c1…

Drug data preprocessing

Model 1