DrugVLABTM은 chemical space, genetic space, disease space를 정확하고 효율적으로 탐색하여 표적단백질 및 후보물질을 선별합니다.
수많은 논문으로 검증된 standalone tool들과 다양한 needs를 모아서 만든 pre-built packages를 통해 약물 개발 단계를 최적화합니다.

Standalone Tools

Target Discovery

Biomarker discovery

  • Biomarkers in predicting lymph node metastasis. Scientific Reports, 2021
  • Risk stratification for breast cancer patient. Cancers, 2022
  • GOAT: biomarker for eosinophilic asthma subtype. Bioinformatics, 2023

EHR data analysis or Text mining

  • Inductive document classification. AAAI, 2022
  • TM-HGNN: patient-level representation learning. ACL, 2023
  • GraphT5: molecular graph-language modeling. 2024

Lead Discovery

Lead identification

  • ChemMining: data-mining with chemical similarity. CSBJ, 2023
  • DiSCO: molecular conformer optimization. AAAI, 2024
  • MSA-VF: virulence factor prediction. 2024

Drug target interaction

  • EnsDTI: ensemble DTI prediction. 2024
  • Diffusion-based molecular docking. 2024
  • ATP binding site prediction. 2024

Lead Optimization

Lead optimization with graph deep learning

  • SPGP: structure prototype guided graph pooling. NeurIPS GLFrontiers, 2022
  • Improved OOD generalization in graph. CVPR, 2024
  • ChemGen: generative AI models for chemical design. 2024

Drug combination 

  • MADC: drug synergy. BIB, 2023
  • Drug-drug interaction. ICLR, 2024
  • Drug relational learning. 2024



  • SSM: liver toxicity. iScience, 2023
  • SIDER: side effect. IEEE JBHI, 2024
  • MDTR: quantifying liver toxicity. 2024


  • Molecular property: multi-task aware learnable prototypes on few shot learning, 2024

Animal model

  • In-silico experimental system. Methods, 2018
  • PULSAR: in-silico virtual knockout system. 2024


Drug response

  • DRPreter: interpretable drug resp. Int J Mol Sci., 2022
  • NetGP: drug resp. with target data. BIB, 2023
  • CSG2A: drug resp. by condition-specific perturbation. ISMB, 2024
  • DrugPT-Net: drug perturbation guided visible neural network for drug resp. 2024

Drug repurposing

  • DREAMwalk. Nature Comms., 2023

Drug approval/withdrawal

  • Drug withdrawal pred. IEEE BigComp, 2022
  • ChemAP: drug approval pred. 2024


데이터 마이닝 기반 화합물 데이터 베이스 탐색으로 선도물질 발굴 가속화

ChemMining™ efficiently explores the chemical space, identifying high-efficacy compounds and providing priority rankings utilizing Deep Learning, Network Exploration, and Cheminformatics. Our innovative approach allows for the discovery of compounds with unique scaffolds, enhancing patentability and diversity.

  • Used Cases Discovery of hit compounds for small molecule kinase inhibitors. Identification of hit compounds for pharmacological chaperones.


화합물의 functional fragments 발굴을 통한 3차원 단백질 구조 기반 선도물질 생성 가속화

ChemGen™ innovates compound design using protein 3D structures and experimental data. Our Generative AI models create molecules binding to protein pockets. With Network Propagation technology, we efficiently select high-efficacy, drug-like compounds with prioritized precisions.

  • Used Cases Optimization of pharmacological chaperone lead compounds. Development of optimal compound formulations.


Drug Response Platform

ChemResponse™ predicts drug responses by analyzing the mechanism of action from various perspectives. Our innovative approach identifies crucial gene interactions and pathways. Also, we simulate the effects of compound perturbations starting from target genes, pinpointing driver genes crucial to the biological mechanisms of the compound.

  • Used Cases Interpreting mode of action of compounds. Stratification of samples based on drug responsiveness.


생체 내 데이터 기반 인체 내 독성 및 부작용 예측

ChemTox™ analyzes in vivo toxicity and side effects based on transcriptome data and compounds structures. We identify crucial toxicophores, predict toxicity via CYP and hERG analysis, and offer insights into six hepatotoxicity mechanisms. Also, we forecast compound side effect frequencies based on molecular structure, similarity, and target protein data.

  • Used Cases Prediction of kinase off-target effects.


생물의학적 지식 네트워크 기반 약물 재창출 및 복합제 예측

ChemTune™ utilizes biomedical knowledge networks for knowledge drug repurposing and synergy analysis. Our biomedical knowledge graph connects drug-disease spaces with Protein-Protein Interaction (PPI) networks. Skip-Gram encoder models analyze relationships, and ‘Teleportation’ in random walks ensures unbiased exploration.

  • Used Cases Prediction of small molecule inhibitor combinations.


3차원 구조정보 기반 PROTAC 링커 설계 및 ADME 예측

ChemProtac™ elevates the potential of candidate inhibitors by expertly designing tailored PROTAC solutions. By integrating existing assay data with advanced protein structure prediction and docking simulation models, our model refines the search for optimal linkers in the development and evaluation of de novo PROTAC leads.

  • Used Cases Explore innovative PROTAC linker designs. Predicting protein inhibition efficacy for existing PROTACs.

Customized Packages

Fully customized AI service tailored specifically for your needs