Development of Machine Learning-Based Methods for Predicting Small Molecules Binding to Specific RNA Targets
We are developing technology to predict small molecules that bind to target RNAs by training molecular classifiers using compound screening results as labels (supervised data) and molecular descriptors as features. These classifiers are then applied to unknown libraries to predict potential binders for the target RNA. Only descriptor information is used, not chemical structures, so there is no risk of structural information leakage.
Our current focus is on enhancing the performance of Focused Library Prediction technology for specific targets. By "Focused Library" here, we do not mean a general RNA-focused library commonly used in the pharmaceutical industry, but a focused library for each specific RNA or DNA target. Since the training data comprises screening results for specific targets, it enables the prediction of a focused library tailored to individual targets.
The specific target does not have to be nucleic acids. Focused libraries can be constructed for any type of target. If you are interested, please feel free to contact us.
