Quantitative Medicine’s Computational Research Engine™ (CoRE™) speeds drug discovery by using a proprietary combination of machine learning methods to construct a comprehensive predictive model. The model encompasses a very large number of compounds on many targets, measured across numerous experimental modalities. The CoRE™ is unlike other drug discovery methods such as Quantitative Structure Activity Relationship (QSAR), which focuses on predictions for a single protein target, or polypharmacological approaches that generally focus on relatively few targets.
In contrast to most current research strategies, the CoRE™ predicts the results of experiments across numerous, diverse targets and experimental modalities concurrently. The system’s comprehensive model effectively predicts:
- The effects of multiple drug compounds on numerous, diverse targets that mediate disease.
- The combinatorial interactions of drug compounds with other compounds, biomolecules or pathways that could cause adverse effects.
- The impact of numerous attributes such as dosage, concentration and duration
Skillfully developed models can make very accurate predictions.
Predictive analytics is the fast growing applied science of developing predictive models. The models are data derived and usually support decision making. Building upon techniques from traditional statistics, operations management, and artificial intelligence, the CoRE’s predictive analytics use active machine learning methods to structure and find patterns in data. The system incorporates this newly acquired knowledge into models that can make predictions about unknown information.
The CoRE™ uses predictive analytics to develop models of the relationships in an experimental space of biologically relevant data. Because of the complexity and size of these datasets, the system uses sophisticated methods to understand characteristics of the data – such as patterns, noise, and missing data. Using a combination of the most advanced technologies and patent-pending proprietary methodology, the active machine learning models in the CoRE™ are able to efficiently direct experimentation to understand the relationships that are most relevant to a drug discovery campaign.