Developing FFLUX to predictively model short peptide fibrillation

点击翻译按钮获取中文摘要

- the research questions the project is trying to address/the objectives of the project; In past decade the CoI's group has developed a platform for the design of peptide hydrogels based on a family of short amphipathic peptides (typically 8 to 12 amino acids long, each amino acid expressed below by a letter code, e.g. V=valine) with alternating hydrophilic and hydrophobic residues. The overall objective is to systematically test and further develop the novel FFLUX-REG method of the PI's group in both the interpretation and prediction of beta-sheet fibrillation. This computational tool will then guide peptide design. In particular, we will: 1. Investigate with FFLUX-REG the octapeptides FEFKFEFK and AEAKAEAK, and obtain a correct and deep understanding as to why the former is a strong beta-sheet former while the latter is not.2. Understand and control conformational space when swapping A with F: e.g. AEFKAEFK, FEAKFEAK to name a few.3. Understand the role of hydrophobic (replacement of F with L, E, I and V) and hydrophilic (replacement of E by D, and K by R) amino acids.- the approach that will be taken to answer these questions (what the student will actually be doing); The associated scientific protocol and methodology is called FFLUX, which is a completely new force field, designed by novel principles and encoded as a software package. This makes the project truly ground breaking and ambitious because it goes far beyond using standard force fields such as AMBER. The next-generation in-house force field FFLUX is much more realistic than a point-charge based force field such as AMBER. Moreover, FFLUX "sees the electrons" and is hence closer to the underlying quantum mechanics that ultimately governs the behaviour of all matter. FFLUX also introduces multipole moments, which is essential for accurate electrostatics. There is a modern and accurate energy partitioning method called Interacting Quantum Atoms (IQA), which offers a step change in the rigour of atomistic energy analysis. IQA is a parameter-free method that is intuitive but, at the same time, very close to the quantum mechanical character of atoms themselves. There are three types of activities for the student:- development of methodology and coding: (i) improvement of Gaussian Process Regression machine learning (ML) with the goal of training for water clusters, (ii) so-called "knitting" procedure where ML models of smaller systems make predictions for atomic properties in large systems, and (iii) incorporation into FFLUX of post-Hartree-Fock wavefunctions in terms of dispersion energies.- Running of DL_FFLUX on octapeptides in aqueous solution and analysing trajectories.- Running of the in-house REG.py code in order to rank individual energy contributions with an eye on explaining the behaviour of total energy profiles.- the novel engineering and/or physical sciences content of the research (the science that places it within EPSRC's remit). Overall the project fits under two sub-portfolios of the Theme Physical Sciences, namely "Computational and Theoretical", and "Condensed Matter: Electronic Structure". This project also resorts under the Chemical Sciences Grand Challenge of Directed Assembly of Extended Structures with Targeted Properties (DAESTP). Finally, there is a strong Machine Learning component to this project, and thus overlap with Artificial Intelligence. Given the longer term impact on peptides, this proposal will fit the priority Chemical Biology and Biological Chemistry. Furthermore, in the very long term, this proposal is relevant for the grand challenge Healthcare Technologies. There is a strong Machine Learning component to this project, and thus overlap with Artificial Intelligence.