Google DeepMind推出了AlphaGenome Atlas平台，这是一个包含人类基因组中DNA变化预测图谱的AI工具12。该平台汇集了针对90亿个单核苷酸变异的预测数据2，数据库规模达1 petabyte，是AlphaFold数据库的30倍以上2。平台通过预计算AlphaGenome模型的预测结果，为研究人员提供了可直观访问的资源库，用于理解遗传变异如何影响分子生物学2。
实际应用与研究进展
该平台已在多项研究中取得具体成果。在对UK Biobank超54000名参与者的全基因组数据研究中，研究人员发现了22%更多的非编码遗传关联2。Broad Institute的研究人员利用平台新引入的AVI（AlphaGenome Variant Impact）评分——该评分综合了AlphaGenome和AlphaMissense两个模型的预测——发现了与癫痫脑病相关的DNM1基因变异2。平台覆盖数百种人类和小鼠细胞类型和组织2。
开放与前景
AlphaGenome Atlas已面向学术研究免费开放，商业用途可通过Google Cloud获取2。研究人员认为这一平台可能加速科学研究，并为新疾病治疗方法的开发提供可能1。
Google DeepMind has released AlphaGenome Atlas, a comprehensive computational resource containing predictive data for 9 billion single nucleotide variations across the human genome 2. The platform integrates predictions from AlphaGenome and AlphaMissense models into an AVI (AlphaGenome Variant Impact) scoring system, enabling researchers to understand how genetic variations affect molecular biology 2. With a database size of 1 petabyte—more than 30 times larger than the AlphaFold database—the Atlas spans hundreds of human and mouse cell types and tissues 2.
The platform has already yielded concrete research advances. Researchers analyzing data from over 54,000 UK Biobank participants discovered 22 percent more non-coding genetic associations than previously identified 2. Scientists at the Broad Institute used the AVI scoring system to identify genetic variants in the DNM1 gene linked to epileptic encephalopathy 2. These findings demonstrate the Atlas's potential to accelerate scientific discovery and pave the way for developing new disease treatments 2.
AlphaGenome Atlas is currently available free of charge for academic research, with commercial access available through Google Cloud 2.
评论
还没有评论，欢迎留下第一条。