Products and Services

Home > Products and Services > Products > CzeekS

CzeekS (for Virtual Screening)High speed compound screening system using CGBVS

Our CGBVS technology enables the clients to use in-house data

In Chemical Genomics-Based Virtual Screening (CGBVS), active compounds are predicted from the binding patterns obtained from the interaction (chemogenomics data) of proteins (biological space) and compounds (chemical space). The use of CzeekS with the clients' in-house assay data enables creation of more refined models resulting to faster and more precise screening for drug candidates.

● Employs interaction machine-learning method (CGBVS)
CGBVS is implemented via a command line interface (cli); enables high speed and highly accurate in silico compound screening.
● Enables multi-target screening
Scoring against multiple protein targets can be done taking compound selectivity into account.
● Enables search for the target protein of a particular compound
Scoring against all the proteins present in a selected model can be implemented in a compound-by-compound basis allowing the search for the target protein.
● Includes a line-up of various prediction models
In addition to standard models (GPCR, Kinase, Ion channel, Transporter, Nuclear receptor, Protease), models focusing on subfamilies are also included. Models can be chosen according to the client's specific needs.
● Enables creation of new or refinement of existing models through the addition of client's data
Prediction models can be refined through the addition client's assay data. Prediction accuracy of the learning model is expected to increase by the addition of client's data.
● Supports multi-core CPUs (OpenMP Parallelization)
Running CzeekS in systems with multi-core CPUs enables much faster CGBVS calculations.

Technology Infrastructure

CGBVS (Chemical Genomics Based Virtual Screening ) is a technology developed at the research lab of Professor Yasushi Okuno at Kyoto University. This patented technology, that boasts of high speed and high prediction accuracy, is licensed solely to Kyoto Constella Technologies.

Prediction Models

Standard Models (ver. 3) New

Model Target proteins Interactions Details (Proteins and protein groups)
GPCR 244 339,239 Class Aα, Class Aβ, Class Aδ, Class Aγ, Class B, Class C
Kinase 412 200.201 AGC, CAMK, CMGC, STE, TK, TKL, others
Ion channel 218 145,929 Voltage-gated, Ligand-gated, others
Transporter 151 73,669 Electrochemical, ATPase, ATP-binding cassette, others
Nuclear
receptor
41 130,042 NR1, NR2, NR3, NR4, NR5
Protease 235 154,243 Endopeptidase, Exopeptidase

We can create customized models according to your needs. Please contact us for more details.

ChEMBL22 Models New

Model Target proteins Interactions Details (Proteins and protein groups)
GPCR 216 141,360 Class Aα, Class Aβ, Class Aδ, Class Aγ, Class B, Class C
Kinase 404 139,338 AGC, CAMK, CMGC, STE, TK, TKL, others
Ion channel 146 43,317 Voltage-gated, Ligand-gated, others
Transporter 87 20,083 Electrochemical, ATPase, ATP-binding cassette, others
Nuclear
receptor
40 27,265 NR1, NR2, NR3, NR4, NR5
Protease 203 86,944 Endopeptidase, Exopeptidase

ChEMBL models are created based on data obtained from ChEMBL database (Release 22) and are offered for free. Our Standard models (paid models) above, generally, have greater number of interaction data utilized during model creation, as well as, greater number of target protein coverage than the ChEMBL models. Please browse the protein lists in the next section to check the differences.

Downloads

Includes license and prediction model pricing scheme

Includes instructions on how to perform calculations in the Czeek system