{
  "version": 1,
  "event_id": "evt_5684cfda4eaed05a",
  "url": "https://xiyu.news/events/evt_5684cfda4eaed05a/",
  "json": "https://xiyu.news/api/events/evt_5684cfda4eaed05a.json",
  "type": "product_release",
  "status": "monitoring",
  "category": "technology",
  "title": {
    "zh": "谷歌DeepMind发布AlphaGenome Atlas：绘制人类DNA所有可能字母变化的AI预测图谱",
    "en": "Google DeepMind Releases AlphaGenome Atlas"
  },
  "current_state": {
    "zh": "谷歌DeepMind推出AlphaGenome Atlas，这是一个AI驱动的预测图谱，覆盖人类基因组中每一种可能的单碱基DNA变化。该可检索数据库收录了约90亿个单核苷酸变异（SNV）的分子效应预测和AVI评分，涵盖编码区与非编码区。\n\n与疾病相关的变异大多位于占人类基因组98%的非编码区，而以往工具难以大规模解读这些区域。AlphaGenome Atlas将相关预测转化为可公开检索的资源，有望加速罕见病诊断、变异优先级排序和药物靶点发现。\n\n底层的AlphaGenome模型可输入长达1Mb（约100万碱基）的侧翼DNA序列，从而避免了此前工具在上下文长度与预测分辨率之间的取舍。Atlas收录了约90亿个单核苷酸变异，用户无需填写机构信息即可免费在线访问。",
    "en": "Google DeepMind releases AlphaGenome Atlas, a predictive map of human genome function, integrating multiple resources for easier scientific access."
  },
  "first_seen_at": "2026-09-08T16:47:08.756554+00:00",
  "last_updated_at": "2026-09-08T16:47:08.756554+00:00",
  "last_material_change_at": "2026-09-08T16:47:08.756554+00:00",
  "confidence": 0.75,
  "updates_count": 1,
  "sources_count": 1,
  "entities": [
    "alphagenome",
    "atlas",
    "deepmind",
    "google",
    "releases"
  ],
  "identifiers": [],
  "topics": [
    "bioinformatics",
    "deepmind",
    "genomics",
    "research"
  ],
  "updates": [
    {
      "update_id": "upd_e4b9f096d461ab78",
      "event_id": "evt_5684cfda4eaed05a",
      "occurred_at": "2026-09-08T14:55:45Z",
      "published_at": "2026-09-08T14:55:45Z",
      "first_seen_at": "2026-09-08T16:47:08.756554Z",
      "time_precision": "published",
      "update_type": "initial",
      "material_change": true,
      "title_zh": "谷歌DeepMind发布AlphaGenome Atlas：绘制人类DNA所有可能字母变化的AI预测图谱",
      "title_en": "Google DeepMind Releases AlphaGenome Atlas",
      "what_changed_zh": "谷歌DeepMind推出AlphaGenome Atlas，这是一个AI驱动的预测图谱，覆盖人类基因组中每一种可能的单碱基DNA变化。该可检索数据库收录了约90亿个单核苷酸变异（SNV）的分子效应预测和AVI评分，涵盖编码区与非编码区。\n\n与疾病相关的变异大多位于占人类基因组98%的非编码区，而以往工具难以大规模解读这些区域。AlphaGenome Atlas将相关预测转化为可公开检索的资源，有望加速罕见病诊断、变异优先级排序和药物靶点发现。\n\n底层的AlphaGenome模型可输入长达1Mb（约100万碱基）的侧翼DNA序列，从而避免了此前工具在上下文长度与预测分辨率之间的取舍。Atlas收录了约90亿个单核苷酸变异，用户无需填写机构信息即可免费在线访问。",
      "what_changed_en": "Google DeepMind releases AlphaGenome Atlas, a predictive map of human genome function, integrating multiple resources for easier scientific access.",
      "current_state_zh": "谷歌DeepMind推出AlphaGenome Atlas，这是一个AI驱动的预测图谱，覆盖人类基因组中每一种可能的单碱基DNA变化。该可检索数据库收录了约90亿个单核苷酸变异（SNV）的分子效应预测和AVI评分，涵盖编码区与非编码区。\n\n与疾病相关的变异大多位于占人类基因组98%的非编码区，而以往工具难以大规模解读这些区域。AlphaGenome Atlas将相关预测转化为可公开检索的资源，有望加速罕见病诊断、变异优先级排序和药物靶点发现。\n\n底层的AlphaGenome模型可输入长达1Mb（约100万碱基）的侧翼DNA序列，从而避免了此前工具在上下文长度与预测分辨率之间的取舍。Atlas收录了约90亿个单核苷酸变异，用户无需填写机构信息即可免费在线访问。",
      "current_state_en": "Google DeepMind releases AlphaGenome Atlas, a predictive map of human genome function, integrating multiple resources for easier scientific access.",
      "detailed_summary_zh": "Google DeepMind releases AlphaGenome Atlas, a predictive map of human genome function, integrating multiple resources for easier scientific access.",
      "detailed_summary_en": "Google DeepMind releases AlphaGenome Atlas, a predictive map of human genome function, integrating multiple resources for easier scientific access.",
      "background_zh": "人类基因组由约30亿个DNA碱基对组成，单个位点的碱基差异（即单核苷酸变异）与多种疾病和性状相关。以往大多数计算性变异效应预测工具只能处理较短的序列窗口或编码蛋白质的基因区域，对长片段非编码DNA的理解十分有限。AlphaGenome基于DeepMind在科学AI领域的积累，提供一个可以对更长基因组片段进行推理的统一DNA模型。",
      "background_en": "The human genome consists of about 3 billion DNA base pairs, and small differences at individual positions (single-nucleotide variants) are linked to numerous diseases and traits. Most computational variant-effect predictors have been limited to short sequence windows or protein-coding regions, leaving long non-coding DNA poorly understood. AlphaGenome builds on DeepMind's broader AI-for-science work and provides a unified DNA model that can reason over much longer genomic stretches.",
      "community_discussion_zh": "Hacker News读者的反应总体积极，但提出了不少实际问题，例如启动子序列和调控规则是否被完整纳入，以及能否直接上传23andMe等消费级基因检测数据进行分析。还有评论者指出机构信息并非必填项，并分享了一篇Science博客评论，批评该图谱的预测缺乏广泛的实验验证。",
      "community_discussion_en": "Hacker News readers reacted positively but raised practical questions, such as whether promoter sequences and regulatory grammar are fully represented and whether consumer genomes like 23andMe data can be uploaded for analysis. Others pointed out that the affiliation box is optional and shared a Science blog critique noting that the Atlas's predictions lack broad experimental validation.",
      "market_impact_zh": "",
      "market_impact_en": "",
      "importance_score": 8.5,
      "references": [
        {
          "url": "https://news.ycombinator.com/item?id=49611251",
          "title": "Community discussion"
        },
        {
          "url": "https://www.nature.com/articles/s41586-025-10014-0",
          "title": "Advancing regulatory variant effect prediction with AlphaGenome | Nature"
        },
        {
          "url": "https://deepmind.google/blog/alphagenome-atlas-a-predictive-map-of-every-possible-dna-letter-change-in-the-human-genome/",
          "title": "AlphaGenome Atlas: Molecular predictions for 9 Billion human DNA..."
        }
      ],
      "confidence": 0.75,
      "story_ids": [
        "hackernews:story:49611251"
      ],
      "sources": [
        {
          "url": "https://blog.google/innovation-and-ai/models-and-research/google-deepmind/alphagenome-atlas/",
          "label": "utiiiD",
          "source_type": "hackernews",
          "official": false
        }
      ]
    }
  ]
}