Explaining decisions of graph convolutional neural networks: patient-specific molecular subnetworks responsible for metastasis prediction in breast cancer.
Explaining decisions of graph convolutional neural networks: patient-specific molecular subnetworks responsible for metastasis prediction in breast cancer.
复制标题
DOI：
10.1186/s13073-021-00845-7
复制
发表时间：
2021-03-11
期刊：
影响因子：
12.3
通讯作者：
Beißbarth T
中科院分区：
文献类型：
--
作者：
Chereda H;Bleckmann A;Menck K;Perera-Bel J;Stegmaier P;Auer F;Kramer F;Leha A;Beißbarth T
当代的深度学习方法在各种复杂的预测任务中表现出了尖端的性能。尽管如此，深度学习在医疗保健领域的应用仍然有限，因为深度学习方法通常被认为是不可解释的黑箱模型。然而，机器学习社区最近对解释深度学习技术的数据点特定决策的可解释性方法进行了详细阐述。我们认为，这种解释可以通过解释患者特定的预测来帮助个性化精准医疗决策的需求。分层相关传播（LRP）是一种解释深度学习方法决策的技术。它被广泛用于解释应用于图像数据的卷积神经网络（CNN）。最近，CNN开始向非欧几里德域（如图）扩展。分子网络通常表示为详细描述分子之间相互作用的图。基因表达数据可以被分配到这些图的顶点。换句话说，可以通过利用分子网络信息作为先验知识来构造基因表达数据。Graph-CNN可以应用于结构化的基因表达数据，例如预测乳腺癌的转移事件。因此，需要说明分子网络的哪一部分与预测事件相关，例如，癌症的远处转移。我们扩展了LRP的过程，使其可用于Graph-CNN，并在大型乳腺癌数据集上测试了其适用性。我们提出了图分层相关传播（GLRP）作为一种新的方法来解释图形CNN所做的决定。我们展示了一个健全的检查开发GLRP的手写数字数据集，然后将该方法应用于基因表达数据。我们表明，GLRP提供了患者特异性分子子网络，这些分子子网络在很大程度上与临床知识一致，并识别了肿瘤进展的常见和新颖的潜在药物驱动因素。所开发的方法可能在解释个体患者水平上不同组学数据和先验知识分子网络的背景下的分类结果方面非常有用，例如在精准医学方法或分子肿瘤委员会中。在线版本包含补充材料，可在（10.1186/s13073-021-00845-7）获得。
Contemporary deep learning approaches show cutting-edge performance in a variety of complex prediction tasks. Nonetheless, the application of deep learning in healthcare remains limited since deep learning methods are often considered as non-interpretable black-box models. However, the machine learning community made recent elaborations on interpretability methods explaining data point-specific decisions of deep learning techniques. We believe that such explanations can assist the need in personalized precision medicine decisions via explaining patient-specific predictions. Layer-wise Relevance Propagation (LRP) is a technique to explain decisions of deep learning methods. It is widely used to interpret Convolutional Neural Networks (CNNs) applied on image data. Recently, CNNs started to extend towards non-Euclidean domains like graphs. Molecular networks are commonly represented as graphs detailing interactions between molecules. Gene expression data can be assigned to the vertices of these graphs. In other words, gene expression data can be structured by utilizing molecular network information as prior knowledge. Graph-CNNs can be applied to structured gene expression data, for example, to predict metastatic events in breast cancer. Therefore, there is a need for explanations showing which part of a molecular network is relevant for predicting an event, e.g., distant metastasis in cancer, for each individual patient. We extended the procedure of LRP to make it available for Graph-CNN and tested its applicability on a large breast cancer dataset. We present Graph Layer-wise Relevance Propagation (GLRP) as a new method to explain the decisions made by Graph-CNNs. We demonstrate a sanity check of the developed GLRP on a hand-written digits dataset and then apply the method on gene expression data. We show that GLRP provides patient-specific molecular subnetworks that largely agree with clinical knowledge and identify common as well as novel, and potentially druggable, drivers of tumor progression. The developed method could be potentially highly useful on interpreting classification results in the context of different omics data and prior knowledge molecular networks on the individual patient level, as for example in precision medicine approaches or a molecular tumor board. The online version contains supplementary material available at (10.1186/s13073-021-00845-7).
登录
查看更多内容
影响因子：
--
作者：
Fisher, RA
通讯作者：
Fisher, RA
影响因子：
5.8
作者：
Chen, Li;Liu, Han;Chen, Jun
通讯作者：
Chen, Jun
影响因子：
6.4
作者：
Banerjee K;Resat H
通讯作者：
Resat H
影响因子：
2.1
作者：
Irizarry, RA;Hobbs, B;Speed, TP
通讯作者：
Speed, TP
影响因子：
5.8
作者：
Gendoo, Deena M. A.;Ratanasirigulchai, Natchar;Haibe-Kains, Benjamin
通讯作者：
Haibe-Kains, Benjamin