Precise prediction of phase-separation key residues by machine learning

Precise prediction of phase-separation key residues by machine learning

26 March 2024 | Jun Sun, Jiale Qu, Cai Zhao, Xinyao Zhang, Xinyu Liu, Jia Wang, Chao Wei, Xinyi Liu, Mulan Wang, Pengguihang Zeng, Xiuxiao Tang, Xiaoru Ling, Li Qing, Shaoshuai Jiang, Jiahao Chen, Tara S. R. Chen, Yalan Kuang, Jinhang Gao, Xiaoxi Zeng, Dongfeng Huang, Yong Yuan, Lili Fan, Haopeng Yu, Junjun Ding
The study introduces PSPhunter, a machine learning algorithm designed to predict phase-separating proteins and identify key residues that significantly impact phase separation. The algorithm integrates sequence and functional features, including word2vec, Position-Specific Scoring Matrix (PSSM), Hidden Markov Model (HMM), and various protein properties. PSPhunter was validated through in vitro and in vivo experiments, demonstrating its ability to accurately predict phase-separating proteins and key residues. Key residues identified by PSPhunter were found to be enriched in glycine and proline, which play crucial roles as spacers in phase separation. Pathogenic mutations, particularly those in key residues, were shown to disrupt phase separation more significantly than neutral mutations. The study also highlights the importance of phase separation in transcriptional control, cell fate transitions, and disease development, with specific mutations in key residues affecting tumor cell migration and growth. PSPhunter provides a robust tool for understanding and manipulating phase separation, offering insights into the molecular mechanisms underlying various diseases.The study introduces PSPhunter, a machine learning algorithm designed to predict phase-separating proteins and identify key residues that significantly impact phase separation. The algorithm integrates sequence and functional features, including word2vec, Position-Specific Scoring Matrix (PSSM), Hidden Markov Model (HMM), and various protein properties. PSPhunter was validated through in vitro and in vivo experiments, demonstrating its ability to accurately predict phase-separating proteins and key residues. Key residues identified by PSPhunter were found to be enriched in glycine and proline, which play crucial roles as spacers in phase separation. Pathogenic mutations, particularly those in key residues, were shown to disrupt phase separation more significantly than neutral mutations. The study also highlights the importance of phase separation in transcriptional control, cell fate transitions, and disease development, with specific mutations in key residues affecting tumor cell migration and growth. PSPhunter provides a robust tool for understanding and manipulating phase separation, offering insights into the molecular mechanisms underlying various diseases.
Reach us at info@study.space
Understanding Precise prediction of phase-separation key residues by machine learning