Advancing capability in high performance protein structure and function prediction through optimisation of IntFOLD

点击翻译按钮获取中文摘要

One of the major challenges in biology is to understand how proteins fold up into the different shapes that are specified by their sequences of amino acid building blocks. If we know how proteins fold then we can understand what they do and how they work together as the fundamental molecular machines in all living systems. Our research aims to improve our ability to understand protein structures and how they function. This information can be used to help us tackle a wide range of urgent problems, such as, securing future food supplies, producing new medicines and sources of energy, and ensuring healthier people, plants and animals. Proteins are the most important components of every single living cell and they come in thousands of different shapes and sizes. Genes contain the code for making the many different protein molecules. We have very efficient machines for analysing genes and collecting genetic sequence code. We have already collected the genetic sequences for thousands of living things, from bacteria to plants and animals, but there are still many more to investigate. The amount of available genetic information is increasing at an ever faster rate and we are making strides to decode this information to understand what the encoded proteins do.There are several different types of experiments that we can do to find out the shapes or structures of proteins. Unfortunately, doing an experiment to find out the structure of just one protein can take many years and it can be very expensive. This means that we now have large knowledge gaps with missing information about what proteins look like and how they work together. In order to make full use of the genetic information that we are collecting, we need to be able to close these gaps in our knowledge and complete the puzzle.Fortunately, we have developed our computer software system, called IntFOLD, to model the structures of proteins, which is many times faster and cheaper than physical experiments. The IntFOLD software makes use of our existing knowledge of protein sequences and structures to help fill in the missing information about new sequences. By learning from what we already know, the software can make predictions about the shapes of the new proteins. We can then build virtual models of the molecules and see where all of the atoms are likely to be in three dimensions. We can then better understand how the molecules combine together to form biological machines.This transformative project is about the major enhancement of our IntFOLD software, making it even more useful and promoting it to more biologists in the UK and around the world. The software has already been used hundreds of thousands of times by thousands of researchers worldwide. The models produced by IntFOLD have helped new research into molecular mechanisms, diseases and the evolution of proteins across all kingdoms of life. We now need to improve our IntFOLD software to make the models more precise, which will improve their usefulness further. We also need to include more predictions about how proteins assemble, which will improve our understanding of their functions. To effect this step change, we will need to employ a dedicated post doctoral researcher to assist in the development of the new IntFOLD, as well as to provide its availability to researchers worldwide. Computer speed and capacity is of the essence to keep up with the growth in demand, so we are also requesting funding to keep our hardware up to date.