撰文丨王聪 编辑丨王多鱼 排版丨水成文 多囊卵巢综合征 （PCOS） 是育龄女性中最常见的内分泌紊乱疾病，具有高度遗传性，但我们对其多基因结构仍知之甚少。 2025 年 11 月 4 日，中国科学院脑科学与智能技术卓越创新中心 师咏勇 团队联合山东大学/ 生殖医学与子代健康全国重点实验室 赵涵 教授、 青岛大学/青岛大 学附属医院 李志强 教授、安徽医科大学 曹云霞 教授等，在 Nature Genetics 期刊发表了题为： Multi-ancestry genome-wide association analyses of polycystic ovary syndrome 的研究论文。 在这项最新研究中，研究团队对 12419 名中国 多囊卵巢综合征 （PCOS） 患者和 34235 名对照者进行了全基因组关联研究 （GWAS） ，随后与多达 13773 名欧洲病例和 411088 名对照者进行了多祖先荟萃分析，进而确定了 94 个独立遗传位点，其中 73 个此前未被报道。 该研究发现，尽管进化压力不同，但中国和欧洲血统显示出大量的遗传重叠。功能注释发现， AMH 基因上的错义突变 rs10407022 ...

Core Insights - A new computational tool named TWAVE has been developed by a team from Northwestern University, utilizing generative AI to extract key information from limited gene expression data and identify multi-gene combinations behind complex diseases [1][2] - The TWAVE model simulates gene expression under healthy and diseased states, linking changes in gene activity to phenotypic variations, and accurately pinpointing key gene changes that may trigger cellular state transitions [1][2] Group 1: Technology and Methodology - TWAVE focuses on gene expression levels rather than gene sequences, addressing the limitations of traditional genome-wide association studies (GWAS) that primarily identify single genes associated with specific traits [2] - The model was trained using clinical trial data to recognize expression profiles representing healthy or diseased states, enhancing its ability to identify disease-associated gene networks [2] - TWAVE circumvents privacy issues related to gene sequences and inherently incorporates environmental factors, allowing for a more comprehensive understanding of gene-environment interactions [2] Group 2: Applications and Implications - Testing of TWAVE on various complex diseases demonstrated its capability to identify known pathogenic genes and discover new genes overlooked by existing methods [2] - The findings indicate that the same disease may arise from different gene combinations in different populations, providing a theoretical basis for personalized treatment based on individual genetic drivers [2][3] - The advancements in AI within the life sciences are facilitating a deeper understanding of disease mechanisms and supporting early diagnosis and personalized treatment, accelerating the arrival of the precision medicine era [3]

Core Viewpoint - The article discusses a significant study identifying 64 genes that influence brain aging speed and suggests 13 potential anti-aging drugs, utilizing AI models to analyze brain scans and genetic data [1][3]. Research Overview - The study is noted as the largest attempt to determine genetic factors affecting organ aging, with implications for developing new brain anti-aging drugs [3]. - The research aims to identify factors leading to brain aging and explore potential solutions [5]. Methodology - The study uses Brain Age Gap (BAG) as a marker for brain aging, defined as the difference between predicted brain age and actual age [6]. - Data from 29,097 healthy participants in the UK Biobank was used to train seven AI models for brain age estimation [8]. - Validation was conducted using data from 3,227 healthy and 6,637 brain disease subjects, employing various assessment metrics [9][10]. Genetic Analysis - A Genome-Wide Association Study (GWAS) was performed on 31,520 healthy participants to identify genetic variations associated with BAG [11][12]. - The study explored the causal relationship between BAG and 18 brain diseases, finding a significant impact on intelligence [13][14]. Drug Discovery - The research identified 64 druggable genes linked to biological pathways related to brain aging, suggesting that targeting these genes could help combat aging or related diseases [14][15]. - A drug repurposing analysis revealed 466 potential anti-aging drugs, with 29 showing promise in delaying brain aging [17][18]. - Among these, 20 drugs, including Dasatinib and Diclofenac, have been previously noted for their anti-aging potential, with 13 currently undergoing clinical trials [19][20].