Interprotein Corporation 

A Japanese biotech company is seeking an EU partner for AI-based drug discovery

KEY ACTIVITIES Entropy-considered in silico screening of small molecules and AI-based activity prediction
SECTOR: Biotech
KEY WORDS: Protein-protein interaction, ubiquitin-proteasome system, in silico screening, artificial intelligence, activity prediction
TYPE OF ACTIVITIES Research/R&D

Company Activity Description:

  1. Interprotein established a novel platform technology for in silico screening that is applicable to challenging targets including PPIs and gave it a name of INTerprotein’s Engine for New Drug Design (INTENDDTM).
  2. Interprotein newly developed a deep learning method for structural information and activity information.  One of the most significant points in this process is proprietary data-preprocessing method for structural information and resulting unique training data enable efficient and speedy deep learning.
  3. Hit rates of INTENDD-proposed compounds constantly reach the values of 10% or more for PPI targets and the best rate reached up to 47%.  When we validated accuracy of AI-guided INTENDD using Runx1 inhibitor compounds whose co-crystal structure information had not been obtained, “Good Prediction” rate reached 79%, strongly suggesting that AI-guided INTENDD is applicable to compounds without X-ray co-crystal structure information.
 

Type of Partnership Sought:

1. Research cooperation agreement
2. Technical cooperation agreement
3.  License agreement
4. Financial agreement

 

Additional information:

 

Example of collaborative research model with pharmaceutical companies

.

* Each period needed for INTENDD® or AI-guided INTENDD® varies in dependence upon targets.  These two approaches can be conducted separately.  AI-guided INTENDD®-based lead generation/optimization would be performed by several cycles of prediction and assessment.

 

Company website