Large-scale Drug Function Prediction by Integrating QIS D2 and BioSpice

Quantum Intelligence System for Drug Discovery (QIS D) is a unique adaptive learning system designed to predict potential large-scale drug characteristics such as toxicity and efficacy. BioSpice is a set of software tools designed to represent and simulate cellular processes funded by DARPA. We show a QIS D model is successfully trained, tested and validated on experimental data sets for predicting the potential in vivo effects of drug molecules in biological systems. QIS D is interoperable with BioSpice. The workflow and visualization are built-in capabilities for easy-of-use. The integration of QIS D and BioSpice draw on diversified technologies to deliver unique benefits for simulation and screening of potential drugs and their targets. We show that our approach leverages both structured and unstructured bioinformatics databases such as BioWarehouse and GeneWays in BioSpice to greatly enhance a QIS D model. We show QIS D models data from seven sources for 37,330 chemicals, performs an automatic sequence clustering using 1234 structure fragments, and accurately predict 1829 targets simultaneously.