serviceACISS: In Silico Drug Discovery Support Service

Affinity-Constella In Silico Support

The ACISS (Affinity-Constella In Silico Support) service is a contract research and analysis service jointly provided by Affinity Science, Inc. Ltd., Institute for Theoretical Medicine, Inc. and INTAGE Healthcare, Inc. with the aim of supporting computational science with regard to drug discovery and other related fields. We provide total support for drug discovery research with excellent cost-effectiveness, including in silico screening, validation by simulation, maintenance of compound databases, and introduction of optimal software systems.

Core Technologies
CGBVS (Chemical Genomics-Based Virtual Screening): Compound-Protein Interaction Machine Learning Method
PBVS (Pharmacophore-Based Virtual Screening): Pharmacophore-based Method
SBVS (Structure-Based Virtual Screening): Docking simulation method
LBVS (Ligand-Based Virtual Screening): Similarity search

CGBVS (Compound-protein interaction machine learning method)

CGBVS is a unique computational method that predicts the interaction of unknown protein-ligand pairs through machine learning of known interaction information. In addition to compound discovery, it can also be applied to target prediction calculations and compound design.

|image scrollImage is scrollable

*CGBVS is a patented technology of Kyoto University that has been put to practical use. In addition to our experience in contract calculations, we have a lot of experience in providing applications and implementing them on the "K computer".

PBVS (Pharmacophore-based Method)

The pharmacophore model, which represents the functional group properties of active compounds in three-dimensional coordinates, is constructed and utilized to search for new compounds. The pharmacophore model can be constructed by either fast and accurate conformation analysis of known active compounds (ligand-based) or by using the active conformation of the ligand-target complex (structure-based).

|image scrollImage is scrollable

SBVS (Docking Simulation Method)

We search for compounds that bind to the active pocket by performing simulations using a three-dimensional structure of the target protein.

|image scrollImage is scrollable

LBVS (Similarity Search)

Search for compounds similar to known active compounds based on molecular structure and physical properties.

Conditions required for each calculation method and the characteristics of the results

Method Required data and conditions Characteristics of expected results
CGBVS: Interaction machine learning method
  • Primary sequence information of the target protein
  • Information on known active and inactive compounds e.g. 100 ligand data with IC50<30μM
  • Information on the known activities of proteins similar to the target protein can also be used effectively.
  • Predict interactions based on pattern recognition through machine learning of vast amounts of known activity data
  • Useful for searching for compounds similar to known structures
  • Effective in searching for new compound structures
  • Can be used for target protein prediction
PBVS: Pharmacophore-based Method
  • For ligand-based calculation: information on several known active compounds
  • For structure-based calculation: crystal structure of target protein and ligand complex (PDB file)
  • Three-dimensional conformational models are created from active compounds and used for screening.
  • Compounds with novel structures may be found.
  • Can be performed even when the target protein is unknown (ligand-based calculation)
SBVS: Docking simulation method
  • 3D structure of target protein (PDB file)
  • Complex of known ligand with target protein is preferable
  • Simulation of ligand binding to the protein pocket is performed
  • Useful when searching for novel compound structures
LBVS: Similarity search method
  • Information about several known ligands
  • Useful even if the target protein is unknown
  • Search is performed based on similarity with known active ligands
  • Useful for searching for peripheral structures of known compounds
  • Not suitable when searching for compounds with novel structures

Comparison of features

Feature CGBVS PBVS SBVS LBVS
3d protein structure Not required Can be used when available Required Not required
Information about known ligands Required to a certain degree Required Required to a certain degree Required
Information regarding ligands of neighboring proteins (protein family) Required to a certain degree Not required Not required Not required
Prediction accuracy
Ability to search for novel structures
Speed and cost
Applicability to compound optimization
Applicability to target prediction ×

(This is a summary of the general characteristics of each method. It may vary depending on the requirements of the study, so please contact us for details.)

Regarding available compound libraries for screening

Possible target compound libraries for screening include the following.

  • Compound libraries owned by the client company (tens to hundreds of thousands of compounds)
  • Commercially available compound libraries (Namiki Shoji Co., Ltd., Kishida Chemical Co., Ltd., etc.)
    (Example of a Namiki Shoji compound libraries)
    → In-stock library (5 million compounds)
    → 3 major suppliers (2 million compounds)
    → Suppliers with short delivery times (more than 2.5 million compounds) - approx 1-1.5 months
    → Virtual library (4 million compounds) with synthetic accessibility >85%

Others

Inquiry

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

Contact Us

SHARE

facebooktwitter

INTAGE Healthcare’s Scope of Service