Novel algorithms in protein folding: quantum computing, deep learning and molecular simulation

我们建议与该领域的工业合作伙伴合作，探索如何将量子计算应用于具有生物医学和制药重要性的问题。学生将驻扎在统计和材料系，并根据需要访问合作伙伴公司。我们建议将重点放在里程碑1上，它将允许探索里程碑2和里程碑3。里程碑1：利用运行在经典超级计算机上的量子计算机仿真器开发新的量子计算算法。应用：生物医学机器学习和优化问题概述如下。里程碑2：展示了使用量子计算机比经典计算机具有更好的优化性能。应用：应用量子优化来搜索系统的最低能态，包括(I)晶格模型和粗略获得的残基-残基势来模拟蛋白质折叠；(Ii)简化的成对原子势用于点配基对接；(Iii)使用容易出错的量子计算机的Boltzmann采样来识别小分子和二肽的低能构象。里程碑3：实现一个简单的量子二进制分类器，用于监督学习，比经典方法更有效。应用：预测(I)抗体序列是否会聚集；(Ii)对于给定的蛋白质靶标，配体是结合体还是非结合体。这个项目将有助于NQIT加强和多样化其与行业的联系，并探索开发新的算法应用于生物医学科学。学生将被期望开发算法，并与牛津蛋白质信息学小组(OPIG)和量子纳米技术理论小组(QuNaT)以及合作公司UCB和Roche.具有领域知识的专家密切合作。该项目属于EPSRC量子技术研究领域。

We propose to explore how quantum computing can be applied to problems of biomedical and pharmaceutical importance, in collaboration with industrial partners in this sector. The student will be based in the Departments of Statistics and Materials, and visit partner companies as needed. We propose to foucs on milestone 1, which will permit exploration of milestones 2 and 3.Milestone 1: develop novel quantum computing algorithms using quantum computer emulators running on classical supercomputers.Applications: biomedical machine learning and optimization problems outlined below.Milestone 2: demonstrate superior optimization performance using quantum computer over classical computers.Application: apply quantum optimzation to search for lowest energy state of a system, including (i) a lattice model and coarse-gained residue-residue potential to simulate protein folding; (ii) a simplified pairwise atomic potential for potein-ligand docking; (iii) identify low energy conformers of small molecules and dipeptides using error-prone quantum computer's Boltzmann sampling.Milestone 3: Implement a simple quantum binary classifier for supervised learning that is more efficient than classical approaches.Application: predicting (i) whether an antibody sequence will aggregate or not; (ii) whether a ligand is a binder or non-binder for a given protein target.This project would help NQIT to strenthen and diversify its ties to industry, and to explore the develop of novel algorithms to applications in biomedical sciences. The student would be expected to develop the algorithums, and to work closely with with experts with domain knowledge in both the Oxford Protein Informatics Group (OPIG) and the Quantum Nanotechnology Theory Group (QuNaT), as well as partnering companies, UCB and Roche.This project falls within EPSRC Quantum Technologies research area.