The Purpose Of Protein Folding Of AI Technology : The Creation Of The Heart Of Molecules To Develop Intelligent Life Sciences

AI is driving intelligent life sciencesforward in a big way, expanding the purpose of protein folding. Professor JinboXu, founder and chief scientist of MoleculeMind, gave a talk on the futuretrend of intelligent life science and the purpose of protein folding withProfessor Gao Wen, academician of the Chinese Academy of Engineering anddirector of Pengcheng Laboratory, and Professor Xiaoliang Xie, director ofFuture Forum and Li Zhaoge Chair Professor of Peking University.

According to Prof. Jinbo Xu, AI-drivenintelligent life sciences will have huge development opportunities in the nextdecade. On the one hand, AI has been proven to be of great value for proteinstructure prediction, and the purpose of protein folding, large moleculeprediction, optimization, and design will be greatly enhanced for treatingdiseases and promoting social production progress. On the other hand, AI willalso be used to predict the relationship between gene variants and diseases toimprove the quality of human life.

Prof. Jinbo Xu is a top computationalbiologist in the field of AI protein, and he has been working in the field ofprotein for more than 20 years, and is known as "the founder of AI proteinfolding technology" in the industry. In 2021, two of the world's top journalsfor scientists, Science and Nature, listed it as one of the top 10 scientificbreakthroughs of the year, and MIT Technology Review also included it in the"Top 10 Breakthrough Technologies of 2022". The MIT Technology Reviewalso included it in its list of the "Top 10 Breakthrough Technologies of2022".

The source of these major breakthroughs isa 2016 research project by Professor Jinbo Xu - the RaptorX-Contact method. Hedemonstrated for the first time in the world that deep learning methods can dramaticallyimprove the accuracy of protein structure prediction, with 90% of the proteinshape correctly predicted and even 50-60% of the high resolution predictionpossible. This has led to a shift in the research paradigm in molecular biologyfrom sequence-based research to structure-based research, which in turn hasfacilitated structure-based drug discovery and design and improved theefficiency of protein design from scratch.

In 2020, the AlphaFold2 algorithm developedby DeepMind, a Google company, made a breakthrough in protein structureprediction, which is also based on this approach.

Xu Jinbo believes that AI will continue tobring great help to the life sciences. On the one hand, the life sciences haveentered the "big data stage", can generate the massive amount of datarequired by AI, on the other hand, the life sciences are very complex, it isdifficult to use simple theories or mathematical formulas to describe thephenomenon of life. AI is good at describing the complex relationship betweenvarious variables.

As the combination of AI with proteinstructure and function prediction and design and the purpose of protein foldingis getting deeper, the application space in related industries is graduallyopening up. For example, the prediction of protein structure and function by AIcan help people understand the pathogenesis of many persistent diseases,including cancer and genetic diseases, and find more accurate treatment paths;at the same time, protein optimization by AI can also greatly improve theefficiency of research and development of large-molecule drugs and reducecosts, and large-molecule drugs are generally more effective than chemicaldrugs and less toxic side effects, and in recent years In recent years, largemolecule drugs have received more and more attention from the medical industrybecause they are generally more effective and less toxic than chemical drugs.In addition to the medical industry, proteins can be used as biocatalysts suchas enzymes and hormones, and have a wide range of applications in food,chemical, energy, environmental engineering and other fields, so theimagination of AI protein design in these fields is also huge.

The integration of"industry-academia-research-application" is also the path ofintelligent life science development that Prof. Jinbo Xu promotes. Recently, heannounced the establishment of the AI protein design platform company"MoleculeMind", and set up a professional team with the world's topcomputational biology experts in China, and cooperate with national lifescience laboratories to carry out scientific research and industrialapplications for AI protein prediction and design. Molecular Heart has alsolaunched MoleculeOS, an AI macromolecule optimization and design platform withindependent intellectual property rights, which uses a data-driven deeplearning approach to help industry experts quickly identify and generate themost suitable proteins, and quickly project laboratory research results toindustrial applications at scale. It can be used for the research and design ofpeptides, antibodies, enzymes and small proteins, making the development oflarge molecule innovative drugs predictable and programmable, and improving theefficiency of the whole drug development process, and can also be flexibly andwidely applied to the optimization and design of proteins in chemistry,materials, industry and agriculture.