<p><span style="color: #464e52; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 18px; white-space: pre-line; float: none; display: inline;">研究人員開發(fā)了一種新的人工智能算法，可以生成類似藥物的分子，并且可以針對特定的藥物靶點進行定制。</span></p> <p>&nbsp;</p> <p><img src="https://info.compassedu.hk/sucai/content/1727667306493/1727667306493.webp" width="780" height="440" /></p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">倫敦國王學院和倫敦帝國理工學院的研究人員開發(fā)了一種新的人工智能算法，可以在幾秒鐘內(nèi)生成數(shù)千個虛擬藥物樣分子。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">這種基于計算機的工具在開發(fā)針對特定藥物靶點的新藥方面可能很有價值。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">虛擬文庫篩選是早期藥物發(fā)現(xiàn)的重要步驟，它涉及使用計算工具搜索現(xiàn)有化合物的數(shù)據(jù)庫，</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">以找到那些結(jié)構(gòu)最有可能與特定藥物靶點結(jié)合的化合物，就像拼圖的兩塊拼圖一樣組合在一起。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">一旦確定了合適的化合物，它們就會在實驗室的細胞和動物模型中進行優(yōu)化和測試，然后再進入臨床試驗。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">但現(xiàn)有虛擬文庫的一個局限性是，對合適化合物的搜索僅限于文庫中已經(jīng)存在的化合物，因此難以識別新的潛在藥物。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">研究人員開發(fā)了一種名為 DrugSynthMC（使用 Monte Carlo 的藥物合成）的新 AI 算法，</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">該算法通過從頭開始每秒生成數(shù)千個類藥物分子的化學結(jié)構(gòu)來擴展藥物庫中的多樣性。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">該方法在 <a style="box-sizing: border-box; color: #0a2d50; font-family: KingsBureauGrotFiveOne, sans-serif; text-decoration-line: underline;" href="https://pubs.acs.org/doi/10.1021/acs.jcim.4c01451" target="_blank" rel="noopener"><em style="box-sizing: border-box;">Journal of Chemical Information and Modeling</em></a> 中進行了概述<em style="box-sizing: border-box;">。</em></p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p><strong><span style="color: #41acbd;">&ldquo;DrugSynthMC 非常高效和快速，因為與許多其他從頭藥物設計工具不同，</span></strong></p> <p>&nbsp;</p> <p><strong><span style="color: #41acbd;">它不依賴于使用訓練數(shù)據(jù)集或神經(jīng)網(wǎng)絡架構(gòu)&rdquo;。</span></strong></p> <p>&nbsp;</p> <p><strong><span style="color: #41acbd;">&mdash;&mdash;國王大學分子生物化學高級講師、該研究的共同高級作者 Filippo Prischi 博士</span></strong></p> <p>&nbsp;</p> <p>&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&ldquo;我們表明，DrugSynthMC 可以擴展可用庫中化合物的化學多樣性，克服現(xiàn)有藥物集合的局限性，&rdquo;Prischi 博士說。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">DrugSynthMC 使用一種稱為蒙特卡洛樹搜索的算法&mdash;&mdash;一種根據(jù)定義的一組操作預測所有可能結(jié)果的數(shù)學技術(shù)。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">在這種情況下，DrugSynthMC 按照一小組指令以簡單的文本格式構(gòu)建分子的化學結(jié)構(gòu)，旨在最大限度地發(fā)揮口服藥物的重要特征。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">該團隊發(fā)現(xiàn)，該算法成功地生成了高比例的易于合成、可溶且無毒的化合物。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">重要的是，DrugSynthMC 可以定制以篩選最有可能與特定生物靶標結(jié)合的分子。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">該團隊認為，人工智能算法未來的主要用途將是識別和優(yōu)化針對與特定疾病相關(guān)的蛋白質(zhì)靶標的化合物。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&ldquo;盡管這是一個相當簡單的算法，但它比任何已經(jīng)測試或發(fā)表的更復雜的算法都要高效得多，</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">并且將在人工智能驅(qū)動的定制治療靶點藥物發(fā)現(xiàn)中非常有用。&ldquo;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">倫敦帝國理工學院癌細胞信號讀者、該研究的共同高級作者 Olivier Pardo 博士說。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">該工具公開可供研究界<a style="box-sizing: border-box; color: #0a2d50; font-family: KingsBureauGrotFiveOne, sans-serif; text-decoration-line: underline;" href="https://github.com/RoucairolMilo/DrugSynthMC" target="_blank" rel="noopener">使用</a>。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">&nbsp;</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">這項工作是與法國巴黎第九大學 PRAIRIE 研究所的 Tristan Cazenave 教授和 Milo Roucairol 合作進行的。</p> <p class="Paragraphstyled__ParagraphStyled-sc-176xsi4-0 jfWPBd" style="box-sizing: border-box; margin-top: 0px; margin-bottom: 1.5rem; white-space: pre-line; font-family: KingsBureauGrotFiveOne, sans-serif; font-size: 1.125rem; line-height: 1.6; color: #464e52;">這項工作部分得到了法國政府的支持。</p>