Virtual High Throughput Screening of small molecules with AI
Discover how to speed up drug discovery using our novel Artificial Intelligence technology for screening of small molecules
The traditional drug discovery process is a high-risk, high-return approach that is costly and lengthy.
Our technology in the field of virtual high-throughput screening (vHTS), ZeptoNet, drastically reduces the costs and time needed for drug discovery and drug repurposing.
Drug discovery - a lengthy and essential activity
New drug discovery is a field that requires vast resources, is time-intensive, and requires an integration of a wide variety of specialized skills in each step of the process. It is estimated that 1 in 10 small molecules projects become candidates for clinical trials, and only about 1 in 10 of those compounds will then pass successfully through all clinical trials and reach the approval stage.
Our technology for drug discovery
Kantify has developed ZeptoNet, a novel, highly performing state-of-the-art solution for in silico small molecule screening. This early-stage technology is highly effective for virtual high-throughput screening (vHTS).
Our novel AI virtual HTS solution enhances the drug discovery process, by speeding up hit discovery, from weeks to hours.
Use of transfer learning
Transfer learning is a mechanism used in deep learning where knowledge from a generic problem is transferred to a domain specific problem. ZeptoNet is therefore engineered as a universal model, which may be optimized towards new and independent HTS trials. As such, instead of having to be trained from scratch - which is a common process in hit discovery efforts - ZeptoNet learns underlying biophysical processes that lead to bioactivity based on a variety of bioassays.
Use of novel featurization of both proteins and compounds
Featurization is the process of turning information into a format that is understandable for machine learning algorithms. Most machine learning algorithms have traditionally been limited to using data in tabular format. ZeptoNet includes a novel featurization pipeline that allows both tabular (e.g. compound properties and fingerprints) and non-tabular data to be included (e.g. the graph structure of the compound or target protein). This data can be used as input for the machine learning algorithm, resulting in increased predictive performance.
Use of active learning
ZeptoNet has been specifically tailored to significantly accelerate active learning. This results in fewer trials and fewer screened compounds to find hits in a compound library.
Discovering new and successful drugs is a hard and time-consuming process. The ZeptoNet helps our clients to :
Reduce the time and cost of drug discovery
Reduce clinical failure rate
Get better results in terms of likely hits
Reduce the number of compounds that need to be physically screened
Open access to research of rare diseases
Search larger compound libraries