Multi-level characteristics recognition of cancer core therapeutic targets and drug screening for a broader patient population.
Multi-level characteristics recognition of cancer core therapeutic targets and drug screening for a broader patient population.
复制标题
DOI：
10.3389/fphar.2023.1280099
复制
发表时间：
2023
影响因子：
5.6
通讯作者：
中科院分区：
文献类型：
--
作者：
简介：针对癌细胞突变的靶向治疗由于其明确的生物学机制和针对特定突变的癌症的高度特异性，在临床应用中引起了人们的关注，包括有限的治疗靶点，较少的患者益处，以及获得性易感性。然而，识别真正致命的癌细胞的合成致死治疗靶点仍然不常见，主要是由于代偿机制。 研究方法：在我们追求在癌症和相应的潜在药物中表现出广泛的合成致死性的核心治疗靶点（CTT）的过程中，我们开发了一种机器学习模型，该模型利用了癌症表征的多个水平和维度。这是通过考虑癌症特异性基因的转录和转录后调控以及构建集成统计和机器学习的模型来实现的。该模型结合了Wilcoxon和Pearson等统计数据以及随机森林。通过WGCNA和网络分析，我们确定了SL网络中作为CTT的枢纽基因。此外，我们还建立了非编码RNA（ncRNA）和药物-靶标相互作用的调控网络。 结果：我们的模型已经发现了7277个潜在的SL相互作用，而WGCNA已经确定了13个基因模块。通过网络分析，我们已经确定了30个CTTs在这些模块中的最高程度。基于这些CTT，我们构建了ncRNA调控和药物靶点的网络。此外，通过将相同的过程应用于肺癌和肾细胞癌，我们已经确定了相应的CTTs和潜在的治疗药物。我们还分析了所有三种癌症的共同治疗靶点。 讨论内容：我们的研究结果在各种维度和组织学数据中具有广泛的适用性，因为我们的模型通过从已知的合成致死基因对中学习多维复杂特征来识别潜在的治疗靶点。统计筛选和网络分析的结合进一步增强了对这些潜在目标的信心。我们的方法提供了新的理论见解和方法论支持的CTTs和药物在不同类型的癌症的识别。
Introduction: Target therapy for cancer cell mutation has brought attention to several challenges in clinical applications, including limited therapeutic targets, less patient benefits, and susceptibility to acquired due to their clear biological mechanisms and high specificity in targeting cancers with specific mutations. However, the identification of truly lethal synthetic lethal therapeutic targets for cancer cells remains uncommon, primarily due to compensatory mechanisms. Methods: In our pursuit of core therapeutic targets (CTTs) that exhibit extensive synthetic lethality in cancer and the corresponding potential drugs, we have developed a machine-learning model that utilizes multiple levels and dimensions of cancer characterization. This is achieved through the consideration of the transcriptional and post-transcriptional regulation of cancer-specific genes and the construction of a model that integrates statistics and machine learning. The model incorporates statistics such as Wilcoxon and Pearson, as well as random forest. Through WGCNA and network analysis, we identify hub genes in the SL network that serve as CTTs. Additionally, we establish regulatory networks for non-coding RNA (ncRNA) and drug-target interactions. Results: Our model has uncovered 7277 potential SL interactions, while WGCNA has identified 13 gene modules. Through network analysis, we have identified 30 CTTs with the highest degree in these modules. Based on these CTTs, we have constructed networks for ncRNA regulation and drug targets. Furthermore, by applying the same process to lung cancer and renal cell carcinoma, we have identified corresponding CTTs and potential therapeutic drugs. We have also analyzed common therapeutic targets among all three cancers. Discussion: The results of our study have broad applicability across various dimensions and histological data, as our model identifies potential therapeutic targets by learning multidimensional complex features from known synthetic lethal gene pairs. The incorporation of statistical screening and network analysis further enhances the confidence in these potential targets. Our approach provides novel theoretical insights and methodological support for the identification of CTTs and drugs in diverse types of cancer.
登录
查看更多内容
影响因子：
64.8
作者：
通讯作者：
--
影响因子：
56.9
作者：
CHO, YJ;GORINA, S;PAVLETICH, NP
通讯作者：
PAVLETICH, NP
DOI：
10.1038/nrc.2017.99
发表时间：
2018-01
期刊：
Nature reviews. Cancer
影响因子：
--
作者：
Anastasiadou E;Jacob LS;Slack FJ
通讯作者：
Slack FJ
LncRNA2Target v2.0：人类和小鼠lncRNA靶基因的综合数据库
影响因子：
14.9
作者：
Cheng L;Wang P;Tian R;Wang S;Guo Q;Luo M;Zhou W;Liu G;Jiang H;Jiang Q
通讯作者：
Jiang Q
影响因子：
28.4
作者：
Bennouna, Jaafar;Hiret, Sandrine;Borg, Christophe
通讯作者：
Borg, Christophe