騰訊雲深平臺取得進展：發佈骨架躍遷新算法

{"type":"doc","content":[{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"近日，騰訊雲深平臺在藥物AI算法研究領域取得新進展。"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"傳統藥物研發存在週期長、費用高和成功率低等特點，新藥研發背後的分子設計需要成千上萬次實驗，準確性亟待提高，使用AI技術能極大改善這一現狀。8月下旬，騰訊AI Lab雲深平臺與成都先導藥物開發股份有限公司（以下簡稱“成都先導”）合作，共同設計完成了業內首個經實驗驗證的骨架躍遷分子生成算法（GraphGMVAE），證明模型能高效並準確地尋找新的候選分子。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"分子設計是新藥研發的重要環節，骨架躍遷方法之於分子研究，就好比用樂高積木建房子，改造建築部件或結構，造出更多樣化的分子“建築”。一般來說，分子設計中常用的傳統骨架躍遷方法需要合成幾十個、甚至上百個分子，才能找不到活性不錯的分子，這一過程需要花費大量人力物力。該算法最少只設計合成了7個分子，即可達到相同的效果。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"該研究結果發表於最新一期美國化學學會雜誌ACS Omega上，爲藥物化學專家設計分子時提供更多啓發，降本增效。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"image","attrs":{"src":"https:\/\/static001.infoq.cn\/resource\/image\/76\/cc\/7665dd300e96484d4bae0782fb3100cc.png","alt":null,"title":"","style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":"","fromPaste":false,"pastePass":false}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":"center","origin":null},"content":[{"type":"text","text":"（圖1. 該項成果發表於行業知名期刊ACS Omega上）"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"骨架躍遷的意義"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"骨架躍遷是一種發現結構新穎化合物的策略。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.infoq.cn\/resource\/image\/e2\/46\/e28777aa13656e5745c96e9ffe778746.png","alt":null,"title":"","style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":"","fromPaste":false,"pastePass":false}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" “骨架”概念廣泛應用於藥物化學和藥物設計中，分析和比較活性化合物與衍生物的核心結構。骨架躍遷，是以已知的活性化合物爲起點，通過改變分子的核心結構，獲得新穎的化學結構。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"找到結構新穎的化合物對新藥研發具有重要意義，也是骨架躍遷的主要目的：（1）在已有的化合物分子結構上，產生新穎的化合物系列，增加藥物研發成功率；（2）替換複雜天然產物的局部結構，產生更具選擇性、更優活性的新穎分子；（3）通過改變分子的骨架，改善分子的藥物代謝動力學性質。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"傳統骨架躍遷算法需要基於現有骨架庫設計新分子。受限於骨架庫大小，新分子在活性和多樣性方面表現不足，而不斷更新骨架庫的過程也耗時耗力。近年來，AI技術在分子生成領域取得進展，在探索更多更廣的化學分子與骨架空間方面展示了更多潛力。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"該算法在骨架提取方面結合了人工規則，更接近人類化學家的角色，不依賴於現有骨架庫的大小規模。在一定程度上，AI在分子生成領域能夠探索更多分子空間。算法學到更符合化學定義的骨架，然後生成更多新穎的化學骨架。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"此外，算法採用高斯混合分佈來區分骨架和側鏈部分，滿足了骨架變化而側鏈保持不變的要求。應用在新藥研發的先導化合物優化階段，GraphGMVAE 算法將有望提升小分子設計效率，從而減少人力以及時間成本。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"騰訊AI Lab致力於AI+醫療探索"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"騰訊AI Lab是騰訊企業級人工智能實驗室，於2017年開始“AI+醫療”探索，不斷拓展和深化研究與應用，涵蓋影像篩查、病理診斷、藥物研發多個領域。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"團隊已有多篇論文入選MICCAI、RSNA，CVPR，AAAI等頂級學術會議，並深度參與及主導多項應用落地，2020年11月，騰訊自研的提升蛋白質結構預測精度的新方法曾登上Nature子刊。與合作伙伴合作研發中國首款獲批進入臨牀應用的智能顯微鏡，發佈AI驅動的藥物研發平臺“雲深”，與鍾南山團隊聯合發佈新冠危重症預測模型等。在產業方面，騰訊與安必平、邁瑞等廠家進行了深入合作，推進AI在顯微領域等廣泛應用。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"成都先導藥物開發股份有限公司是一家從事新藥研發的生物技術公司，總部位於中國成都，在英國劍橋、美國休斯頓設有子公司，並於2020年4月在上海證券交易所科創板掛牌上市（股票名稱：成都先導，股票代碼：688222.SH）。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"成都先導併購了坐落於英國劍橋的Vernalis公司，該公司是FBDD\/SBDD技術的領先者。成都先導爲小分子及核酸新藥發現與優化建立了一個國際領先的，以DNA編碼化合物庫的設計、合成和篩選（DEL），以及基於分子片段和三維結構信息的藥物設計（FBDD\/SBDD）爲核心的技術平臺。目前，公司基於數千種不同的骨架結構，已經完成超過萬億種結構全新、具有多樣性和類藥性的DNA編碼化合物的合成，並且已有多個案例證實了其針對已知生物靶點和新興生物靶點篩選苗頭化合物的能力及有效性。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"未來，該算法將應用於騰訊AI Lab“雲深”平臺助力新藥研發，結合平臺ADMET 預測工具優化先導化合物，使新生成的分子既保持生物活性，還能滿足特定的藥代動力學性質。同時GraphGMVAE還將嘗試用於複雜天然產物的結構改造，尋找化學上更易合成、生物活性和性質更優的苗頭化合物。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong"}],"text":"參考文獻"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"1. Yu, Y. et al. A Novel Scalarized Scaffold Hopping Algorithm with Graph-Based Variational Autoencoder for Discovery of JAK1 Inhibitors. ACS Omega (2021) doi:10.1021\/acsomega.1c03613."}]}]}