[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fpEnHptQFKSqO3ZtKJRDplb16P3bkGo4hYax9Y-5jAtQ":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1409,"科研团队成功利用人工智能蛋白语言模型揭示生命演化奥秘","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">为什么不同生物在适应相似环境时，会独立演化出相似的功能？一项最新研究从蛋白质的“高阶特征”层面揭示了这一生命演化奥秘的重要机制。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这项研究由中国科学院动物研究所邹征廷研究员团队完成，成功利用人工智能领域的蛋白语言模型，揭示了蛋白高阶特征在功能适应性趋同演化中的关键作用，为理解生命演化之谜提供了新视角。相关成果已于近日发表于国际学术期刊《美国国家科学院院刊》。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">“高阶特征”破解传统研究局限\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">趋同演化是自然界中一种引人入胜的现象。蝙蝠与齿鲸本是演化上差异巨大的生物类群，却都独立发展出了通过回声定位感知环境的能力。长久以来，科学家们致力于探索这类表型趋同背后的分子机制。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">传统研究方法主要聚焦于蛋白质序列中单个氨基酸位点的趋同变化。然而，越来越多的证据表明，即使没有明确的位点趋同，同源蛋白仍可能通过高阶结构或理化特征的趋同演化实现功能上的相似性。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">“这就像用不同的砖块砌出结构相似的墙。”研究团队成员解释，“虽然组成的氨基酸可能不同，但蛋白质整体的物理化学性质和结构却可以趋向一致，从而实现相似功能。”\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">人工智能助力科学发现\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">面对这一科学难题，邹征廷团队提出了名为“ACEP”的计算分析框架。该框架的核心创新在于利用了预训练蛋白语言模型。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这类语言模型在海量蛋白序列数据上进行预训练，使其能够捕捉序列中复杂的上下文信息和高阶特征，将蛋白序列转化为包含丰富演化信息的高维嵌入向量。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">“蛋白语言模型是能‘读懂’蛋白质‘语法’的人工智能，它能理解氨基酸序列背后更深层次的蛋白结构功能特征和规律。”研究人员介绍。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">ACEP分析流程包括三个关键步骤：首先计算目标类群同源蛋白嵌入向量的真实距离，然后通过模拟中性演化过程构建背景距离分布，最后基于分布对真实距离进行统计检验，判断是否存在显著的高阶特征趋同信号。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">全面分析验证方法有效性\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">为验证ACEP框架的有效性，研究团队对真实蛋白数据进行了全面的计算分析。在已知的经典案例中，如在回声定位哺乳动物的Prestin蛋白和景天酸代谢植物的PEPC\u002FPPCK蛋白上，ACEP均检测到了显著的高阶特征趋同信号。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">更令人振奋的是，全基因组筛选结果显示，ACEP在蝙蝠与齿鲸中识别出数百个具有趋同信号的候选基因。功能富集分析表明，其中部分基因显著关联“感官感知”等与回声定位密切相关的功能条目。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这些基因不仅包含已知的回声定位基因，还发现了多个新候选基因。部分候选基因还得到了正选择检验的支持，增强了其发生适应性趋同的可信度。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F10\u002Fbaa05e47ce9244409ee3445ed62cd7a2\u002FAA1O3loQ.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cspan class=\"ql-lineHeight-1-75\" style=\"color: rgb(187, 187, 187);\">研究示意图（中国科学院动物研究所提供）\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">推动演化生物学研究范式转变\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">该研究首次系统揭示了蛋白高阶特征趋同是适应性演化的重要机制，突破了传统方法仅关注氨基酸位点趋同的局限。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">“这项工作不仅深化了对生命演化规律的认识，也展示了人工智能技术在解析复杂生物问题方面的强大潜力。我们希望未来能实现人工智能技术在演化生物学中更广泛、有效的应用。”邹征廷研究员表示。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">ACEP框架为在全基因组水平系统挖掘基因的复杂适应性趋同模式提供了新工具。专家认为，这一方法论突破为理解生物适应性演化的分子基础开辟了新方向，并将参与推动演化生物学研究范式的转变。研究成果对生物医学、生态学等领域也具有重要的启示意义。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">研究获得国家自然科学基金和中国科学院战略性先导科技专项等支持。目前，ACEP分析框架的代码已在HuggingFace平台开源共享，供全球科研界使用。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】新华社 （记者胡喆、彭韵佳） \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.msn.cn\u002Fzh-cn\u002F%E7%A7%91%E5%AD%A6\u002F%E7%94%9F%E7%89%A9%E5%AD%A6\u002Far-AA1O3HDd?ocid=BingNewsSerp\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fi9n.cn\u002FHiRdK\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F10\u002F7d521ea7022f433596f852d137ce5e28\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F10\u002Fthumbs\u002F7d521ea7022f433596f852d137ce5e28\u002FAI领域.jpg",0,1,48,"2025-10-08 18:34",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A223feb9e-5f96-4bca-9197-a3b0e34f0997%3A0.wav?Expires=1759923572&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=lzBPuWXC7kwm6ArnepVKySnhn0Q%3D",7509176,"223feb9e-5f96-4bca-9197-a3b0e34f0997","2025-10-08 18:30","Research team successfully revealed the mystery of life evolution using artificial intelligence protein language model","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">Why do different organisms independently evolve similar functions when adapting to similar environments? A new study reveals an important mechanism behind this mystery of life evolution from the perspective of \"higher-order features\" of proteins.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This study was completed by the research team of Dr. Zou Zhengting from the Institute of Zoology, Chinese Academy of Sciences. It successfully utilized artificial intelligence-based protein language models to reveal the key role of protein higher-order features in functionally convergent evolutionary adaptation, providing a new perspective for understanding the mystery of life evolution. The related findings have been recently published in the international academic journal \"Proceedings of the National Academy of Sciences of the United States of America.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">\"Higher-order features\" break through traditional research limitations\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Convergent evolution is an intriguing phenomenon in nature. Bats and toothed whales are vastly different biological groups in terms of evolution, yet both have independently developed the ability to perceive their environment through echolocation. For a long time, scientists have been committed to exploring the molecular mechanisms behind such phenotypic convergence.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Traditional research methods mainly focused on convergent changes at individual amino acid sites in protein sequences. However, increasing evidence shows that even without clear site convergence, homologous proteins may achieve functional similarity through convergent evolution of higher-order structures or physicochemical characteristics.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">\"It's like building walls with different bricks,\" explained a member of the research team. \"Although the amino acids may differ, the overall physical and chemical properties and structure of the protein can tend to be consistent, thus achieving similar functions.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">Artificial intelligence helps scientific discovery\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Faced with this scientific problem, Zou Zhengting's team proposed a computational analysis framework called \"ACEP.\" The core innovation of this framework lies in utilizing pre-trained protein language models.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">These language models are pre-trained on massive protein sequence data, enabling them to capture complex contextual information and higher-order features in the sequences, converting protein sequences into high-dimensional embedding vectors containing rich evolutionary information.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">\"Protein language models are artificial intelligence that can 'read' the 'grammar' of proteins, understanding deeper structural and functional characteristics and patterns behind amino acid sequences,\" the researchers introduced.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The ACEP analysis process includes three key steps: first, calculate the real distance of the embedded vectors of homologous proteins in the target group, then build a background distance distribution by simulating a neutral evolutionary process, and finally perform statistical testing on the real distance based on the distribution to determine whether there is a significant higher-order feature convergence signal.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">Comprehensive analysis verifies the effectiveness of the method\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">To verify the effectiveness of the ACEP framework, the research team conducted comprehensive computational analysis on real protein data. In known classic cases, such as the Prestin protein in echolocating mammals and the PEPC\u002FPPCK protein in CAM plants, ACEP detected significant higher-order feature convergence signals.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">More excitingly, whole-genome screening results showed that ACEP identified hundreds of candidate genes with convergence signals in bats and toothed whales. Functional enrichment analysis indicated that some of these genes were significantly associated with functional categories closely related to echolocation, such as \"sensory perception.\"\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">These genes not only include known echolocation genes but also discovered several new candidate genes. Some of the candidate genes received support from positive selection tests, enhancing the credibility of their adaptive convergence.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F10\u002Fbaa05e47ce9244409ee3445ed62cd7a2\u002FAA1O3loQ.jpg\" width=\"undefined\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cspan class=\"ql-lineHeight-1-75\" style=\"color: rgb(187, 187, 187);\">Research schematic diagram (provided by the Institute of Zoology, Chinese Academy of Sciences)\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\" class=\"ql-lineHeight-1-75\">Promote the transformation of evolutionary biology research paradigms\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This study for the first time systematically revealed that protein higher-order feature convergence is an important mechanism of adaptive evolution, breaking through the limitations of traditional methods focusing only on amino acid site convergence.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">\"This work not only deepens our understanding of the laws of life evolution but also demonstrates the powerful potential of artificial intelligence technology in resolving complex biological problems. We hope to achieve more extensive and effective applications of artificial intelligence technology in evolutionary biology in the future,\" said Dr. Zou Zhengting.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The ACEP framework provides a new tool for systematically mining complex adaptive convergent patterns of genes at the whole-genome level. Experts believe that this methodological breakthrough opens up a new direction for understanding the molecular basis of biological adaptive evolution and will participate in promoting the transformation of evolutionary biology research paradigms. The research findings also have important implications for fields such as biomedicine and ecology.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The research was supported by the National Natural Science Foundation of China and the Strategic Priority Research Program of the Chinese Academy of Sciences. Currently, the code of the ACEP analysis framework has been open-sourced and shared on the HuggingFace platform for global use by the research community.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"color: rgb(187, 187, 187);\">【News Source】Xinhua News (Reporter Hu Zhe, Peng Yunjia) \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.msn.cn\u002Fzh-cn\u002F%E7%A7%91%E5%AD%A6\u002F%E7%94%9B%E7%89%A9%E5%AD%A6\u002Far-AA1O3HDd?ocid=BingNewsSerp\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fi9n.cn\u002FHiRdK\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is reprinted by this website to provide readers with more information and news. The content involved does not constitute investment or consumption advice. If there are any questions about the facts of the article, please verify with the relevant parties. The views of the article are not the views of this website and are for reference only.)\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A98098f98-7175-4234-ace3-ea1ce75559df%3A0.wav?Expires=1774838462&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=CKmN3%2FAc8TPSNbLJt%2BR4%2BeuuQjY%3D","98098f98-7175-4234-ace3-ea1ce75559df",10328136]