Using biological constraints to improve prediction in precision oncology.
Using biological constraints to improve prediction in precision oncology.
复制标题
DOI：
10.1016/j.isci.2023.106108
复制
发表时间：
2023-03-17
期刊：
影响因子：
5.8
通讯作者：
Marchionni, Luigi
中科院分区：
文献类型：
--
作者：
Omar, Mohamed;Dinalankara, Wikum;Mulder, Lotte;Coady, Tendai;Zanettini, Claudio;Imada, Eddie Luidy;Younes, Laurent;Geman, Donald;Marchionni, Luigi
许多基因签名已经通过将机器学习（ML）应用于组学图谱而开发出来，然而，它们的临床实用性往往受到有限的可解释性和不稳定性能的阻碍。在这里，我们展示了在ML方法产生的决策规则中嵌入先验生物知识以构建鲁棒分类器的重要性。我们通过对基因表达数据应用不同的ML算法来预测三种困难的癌症表型：膀胱癌进展为肌肉浸润性疾病，三阴性乳腺癌对新辅助化疗的反应，以及前列腺癌转移进展。我们开发了两组分类器：机械分类器，通过将训练限制在捕获特定生物机制的特征上;不可知分类器，其中训练不使用任何先验生物信息。机械模型有一个类似的或更好的测试性能比他们的不可知论者同行，增强了可解释性。我们的研究结果支持使用生物学限制来开发具有高翻译潜力的强大基因签名。大多数基因签名遭受过度拟合和有限的可解释性在训练中使用已知的生物学可以产生强大的机械分类器我们比较了机械模型与标准不可知模型的性能机械模型往往具有强大的性能，具有增强的可解释性精密医学;癌症;组学;机器学习。
Many gene signatures have been developed by applying machine learning (ML) on omics profiles, however, their clinical utility is often hindered by limited interpretability and unstable performance. Here, we show the importance of embedding prior biological knowledge in the decision rules yielded by ML approaches to build robust classifiers. We tested this by applying different ML algorithms on gene expression data to predict three difficult cancer phenotypes: bladder cancer progression to muscle-invasive disease, response to neoadjuvant chemotherapy in triple-negative breast cancer, and prostate cancer metastatic progression. We developed two sets of classifiers: mechanistic, by restricting the training to features capturing specific biological mechanisms; and agnostic, in which the training did not use any a priori biological information. Mechanistic models had a similar or better testing performance than their agnostic counterparts, with enhanced interpretability. Our findings support the use of biological constraints to develop robust gene signatures with high translational potential. Most gene signatures suffer from overfitting and limited interpretability Using known biology in the training can yield robust mechanistic classifiers We compared the performance of mechanistic models to standard agnostic ones Mechanistic models tend to have robust performance with enhanced interpretability Precision medicine; Cancer; Omics; Machine learning.
登录
查看更多内容
影响因子：
4.5
作者：
Chicco D
通讯作者：
Chicco D
影响因子：
--
作者：
Re A;Corá D;Taverna D;Caselle M
通讯作者：
Caselle M
一种新的AP-1/miR-101调节反馈环路及其在肝癌细胞迁移和侵袭中的意义
影响因子：
14.9
作者：
Liu JJ;Lin XJ;Yang XJ;Zhou L;He S;Zhuang SM;Yang J
通讯作者：
Yang J
影响因子：
30.8
作者：
Bhandari, Vinayak;Hoey, Christianne;Bristow, Robert G.
通讯作者：
Bristow, Robert G.
影响因子：
6.4
作者：
Boormans, Joost L.;Korsten, Hanneke;Trapman, Jan
通讯作者：
Trapman, Jan