An end-to-end approach for protein folding by integrating Cryo-EM maps and sequence evolution

• DOI: 10.1101/2023.11.02.565403
• Publication Date: 2023-11-05T19:30:18Z
• Authors: Pan Li; Liangyue Guo; Haibin Liu; Binghua Liu; Fanhao Meng; Xiaodan Ni; Allen Chunlong Guo
• Keywords: Benchmark (surveying); Sequence (biology); End-to-end principle; Folding (DSP implementation); Cryo-Electron Microscopy; Protein sequencing

Abstract Protein structure modeling is an important but challenging task. Recent breakthroughs in Cryo-EM technology have led to rapid accumulation of density maps, which facilitate scientists determine protein structures it remains time-consuming. Fortunately, artificial intelligence has great potential automating this process. In study, we present SMARTFold, a deep learning prediction model combining sequence alignment features and map features. First, using map, sample representative points along the predicted high confidence areas backbone. Then extract geometric these integrate with our proposed folding model. Extensive experiments confirm that performs best on both single-chain multi-chain benchmark dataset compared state-of-the-art methods, makes reliable tool for atomic determination from maps.