[dt_fancy_title title=”CzeekS” title_align=”left” title_size=”h2″ title_color=”accent” separator_style=”disabled”]

Optimization algorithm-inspired de-novo compound design system

Product Features

  • Ease of use geared towards medicinal chemists
  • Cloud computing service
  • Efficient search of chemical space through the use of optimization algorithms

czeeks_1Click image above to enlarge

[dt_fancy_title title=”Optimization Step” title_align=”left” title_size=”h4″ title_color=”custom” separator_style=”disabled” custom_title_color=”#a81313″]

czeeks_2Fragment-based compound design

  • Compounds are designed through the linkage of fragments based on the synthesis frame. A library of compounds that can be chemically synthesized is created using rules based on RECAP
    RECAP – Retrosynthetic Combinatorial Analysis Procedure
  • Frame setting
    • Compounds are designed through the fitting of fragments to arbitrary frames.
  • Construction of various chemical libraries after functional evaluation using CGBVS
    • Multi-target screening via Chemical Genomics-Based Virtual Screening (CGBVS) enables the design of compounds with very high specificity, as well as, compounds with multiple protein targets.
[dt_fancy_title title=”Chemical Genomics-Based Virtual Screening” title_align=”left” title_size=”h4″ title_color=”custom” separator_style=”disabled” custom_title_color=”#a81313″]

czeeks_3Chemical Genomics-Based Virtual Screening (CGBVS)

  • CzeekS utilizes a machine learning method for virtual screening based on knowledge of chemical-genomics. High-speed virtual screening with high-accuracy is implemented in this system through a command line interface.
  • Compound screening by multi-target prediction
    • Scoring against multiple protein targets enables screening that takes the selectivity of compounds into consideration.
  • Search for each compound’s target protein(s)
    • Scoring of each compound against all the proteins included in the prediction model can be done facilitating the discovery of the target proteins.
  • A lineup of several prediction models
    • In addition to standard prediction models (GPCR, Kinase, Ion channel, Transporter, Nuclear receptor, Protease), models focusing on subfamilies are also available. Clients can choose only the models that suit their purpose.
  • In-house data can be used to create prediction models
    • Clients’ data obtained from in vitro assays can be used to refine existing prediction models. The accuracy of the prediction model is expected to increase through the addition of new data.
  • Compatibility with multicore processors (OpenMP)
    • The ability of CzeekS to utilize multicore processors enables rapid CGBVS simulations.