serviceCzeekS: Rapid compound screening system based on CGBVS

Rapid compound screening system based on CGBVS

Ability to utilize in-house data using the CGBVS method

In Chemical Genomics-Based Virtual Screening (CGBVS*1,2,3), active compounds are predicted based on the binding patterns extracted from the interaction information (chemical genomics information) between proteins (biological space) and compounds (chemical space).
CzeekS enables pharmaceutical companies and research institutes to screen drug candidate compounds with high speed and accuracy using their own accumulated in-house assay data.

*1 Analysis of multiple compound-protein interactions reveals novel bioactive molecules. Mol. Syst. Biol. 7, 472, 2011

*2 Systems biology and systems chemistry: new directions for drug discovery. Chem. Biol. 19(1), 23-8, 2012

*3 Unifying Bioinformatics and Bioinformatics and Computational Modeling, 99-120, 2011

  • Perform screening calculations using compound-protein interaction machine learning method (CGBVS)
    Provides a system to run the CGBVS on the command line. Enables high-speed in silico compound screening with high prediction accuracy.
  • Enables compound screening via multi-target prediction
    Scoring against multiple target proteins allows screening of compounds based on their selectivity.
  • Enables target search for compounds
    For each compound, a score calculation can be performed for all proteins included in the predictive model, which enables search for target proteins.
  • Lineup of various prediction models
    Eight protein families (GPCR, Kinase, Ion channel, Transporter, Nuclear receptor, Protease, Cytochrome P450, PPI) are available as standard predictive models.
  • Ability to create predictive models by adding own data
    Predictive model can be refined by adding own in-house assay data leading to improved prediction accuracy of the machine learning model.
  • Support for multi-core processing (Parallel OpenMP)
    Calculations can be performed faster when CzeekS is used on a machine with multi-core CPU.

Base Technology

CGBVS (Chemical Genomics Based Virtual Screening ) is a computational method developed by Professor Yasushi Okuno of Kyoto University's Graduate School of Pharmaceutical Sciences, which boasts of high speed and high prediction accuracy. Kyoto University has granted INTAGE Healthcare a license for the use the technology in its business.

System License Pricing

Please contacts us.

Predictive Models

The latest ChEMBL predictive model uses alvaDesc as the compound descriptor and PROFEAT 2016 or Multiple Sequence Alignment (MSA) as the protein descriptors. The standard model provided is created using data from the ChEMBL database. In addition to ChEMBL, the PPI model includes data from the TIMBAL database, which is a database dedicated to PPI data.

Model No. of Target proteins No. of training data
GPCR 230 160,941
Kinase 415 159,572
Ion channel 178 59,129
Transporter 120 44,667
Nuclear receptor 41 41,091
Protease 225 118,696
Cytochrome 31 58,658
PPI 53 64,944

*Cytochrome and PPI models are offered since 2019.

Target proteins

List of proteins included in each predictive model.



Ion Channel


Nuclear Receptor


Cytochrome P450



Inquiries about any of the above services can be done through the link below.

Contact Us



INTAGE Healthcare’s Scope of Service