Hybrid Quantum Classical Computing to Aid Computer Aided Drug Discovery

Currently, the computation-heavy part of early-stage drug discovery relies mainly on classical computing and can take from three-five years. There is a need to improve the speed at which drugs can be discovered. Quantum computers can provide an exponential speedup in several kinds of computation problems, like factorisation, in comparison to classical computers. These computers perform computation by manipulating entangled quantum states. The project aims to speed up the calculation of binding affinities in the lead discovery phase of structure-based CADD using quantum computing. By using quantum computing, a speedup from the classical computing approach is expected. This would eventually lower the time taken in the early drug discovery pipeline making the drugs available earlier for pre-clinical trials. Quantum systems are extremely sensitive to atmospheric noise (Linke et al., 2017). Due to the difficulty in achieving highly stable quantum systems Fault-Tolerant Quantum Computing (FTQC) devices are still not achievable. Hence, this project focuses on using a hybrid quantum-classical approach on Noisy-Intermediate Scale Quantum (NISQ) devices. Variational algorithms work efficiently with NISQ devices because of their ability to split tasks between a quantum and classical computer to overall increase efficiency and reduce errors. The Quantum computer works on a short state preparation step while the classical computer provides the quantum computer with the right parameters for state preparation. Currently, for the development of this project, a literature review was conducted around drug design, NISQ devices, and Variational algorithms. Research was done into understanding Variational Quantum Eigensolvers extensively and the entire pipeline of operation. Based on that the algorithm was used along with algorithmic error mitigation techniques.The results saw a good improvement in the accuracy of the results after the error mitigation technique was applied in most cases.