On protocols and measures for the validation of supervised methods for the inference of biological networks.
On protocols and measures for the validation of supervised methods for the inference of biological networks.
复制标题
DOI：
10.3389/fgene.2013.00262
复制
发表时间：
2013-12-03
影响因子：
3.7
通讯作者：
Geurts P
中科院分区：
文献类型：
--
作者：
Schrynemackers M;Küffner R;Geurts P
网络提供了分子生物学知识的自然表示，特别是对基因、蛋白质、药物或疾病等生物实体之间的关系进行建模。由于阐明这些网络所需的努力、成本或缺乏实验，网络推理的计算方法在文献中经常被研究。在本文中，我们研究了监督网络推理的评估。监督推理基于机器学习技术，该技术根据已知交互实体和可能非交互实体的训练样本以及附加测量数据来推断网络。虽然这些方法非常有效，但它们在计算机中的可靠验证提出了挑战，因为预测和验证都需要在相同的部分已知网络的基础上执行。交叉验证技术需要专门适应对象对的分类问题。我们对文献中提出的协议和措施进行严格的审查和评估，并得出如何最好地利用和评估用于网络推理的机器学习技术的具体指南。通过理论思考和计算机实验，我们深入分析了重要因素如何影响性能评估的结果。这些因素包括相互作用实体可用的信息量、生物网络的稀疏性和拓扑结构以及缺乏经过实验验证的非相互作用对。
Networks provide a natural representation of molecular biology knowledge, in particular to model relationships between biological entities such as genes, proteins, drugs, or diseases. Because of the effort, the cost, or the lack of the experiments necessary for the elucidation of these networks, computational approaches for network inference have been frequently investigated in the literature. In this paper, we examine the assessment of supervised network inference. Supervised inference is based on machine learning techniques that infer the network from a training sample of known interacting and possibly non-interacting entities and additional measurement data. While these methods are very effective, their reliable validation in silico poses a challenge, since both prediction and validation need to be performed on the basis of the same partially known network. Cross-validation techniques need to be specifically adapted to classification problems on pairs of objects. We perform a critical review and assessment of protocols and measures proposed in the literature and derive specific guidelines how to best exploit and evaluate machine learning techniques for network inference. Through theoretical considerations and in silico experiments, we analyze in depth how important factors influence the outcome of performance estimation. These factors include the amount of information available for the interacting entities, the sparsity and topology of biological networks, and the lack of experimentally verified non-interacting pairs.
登录
查看更多内容
影响因子：
3
作者：
Hue M;Riffle M;Vert JP;Noble WS
通讯作者：
Noble WS
影响因子：
3
作者：
Lapins M;Wikberg JE
通讯作者：
Wikberg JE
影响因子：
3.7
作者：
Emmert-Streib F
通讯作者：
Emmert-Streib F
影响因子：
3
作者：
Cerulo L;Elkan C;Ceccarelli M
通讯作者：
Ceccarelli M
影响因子：
1.7
作者：
Fluss, R;Faraggi, D;Reiser, B
通讯作者：
Reiser, B