Drug Discovery System
NREC developed an advanced computer vision system that identifies and classifies the behavioral effects of new drug compounds, speeding the work of drug discovery. This drug discovery system is in widespread daily use by the client, Psychogenics, Inc. Tens of thousands of new compounds have been screened so far, with promising new therapies identified for further research and development.
Application
During drug discovery, biotechnology and pharmaceutical companies evaluated chemical compounds to find candidates for new drugs. For drugs that treat neuropsychiatric and neurodegenerative disorders, drug discovery involves observing and documenting the behavior of test subjects. Thousands of compounds must be administered and their behavioral effects observed to find a few good drug candidates.
The drug discovery process is time-consuming, labor intensive, and expensive. Automation increases the number of drug candidates that can be screened, makes behavioral observations more reliable and less biased, and lowers the overall cost. This makes it easier to identify compounds that could potentially be developed into effective drugs.
Capabilities
The SmartCube® observes and classifies the behaviors of individual subjects, such as moving, digging and grooming. It provided a series of challenges to a subject, then compares the subject’s behavioral profile with those from a database. Sophisticated 3D image processing algorithms identify these behaviors.
NREC also developed CognoCube that can handle multiple test subjects and classifies social and behavioral interactions in a group of subjects. It can identify the effect of compounds on the subjects’ cognitive ability, which helps to identify compounds that can be used to treat cognitive impairment. The system uses powerful data mining techniques to convert the recorded behaviors to drug signatures. Essentially, it learns to associate behaviors with chemical compounds.
SmartCube® and CognoCube are more reliable, less subjective and have higher throughput than human observers.
Photos
