Johnson Rob, Kirk Paul, Stumpf Michael P H
Centre for Integrative Systems Biology and Bioinformatics, Department of Life Sciences, Imperial College London, London SW7 2AZ, UK.
Bioinformatics. 2015 Feb 15;31(4):604-5. doi: 10.1093/bioinformatics/btu675. Epub 2014 Oct 16.
Model selection is a fundamental part of the scientific process in systems biology. Given a set of competing hypotheses, we routinely wish to choose the one that best explains the observed data. In the Bayesian framework, models are compared via Bayes factors (the ratio of evidences), where a model's evidence is the support given to the model by the data. A parallel interest is inferring the distribution of the parameters that define a model. Nested sampling is a method for the computation of a model's evidence and the generation of samples from the posterior parameter distribution.
We present a C-based, GPU-accelerated implementation of nested sampling that is designed for biological applications. The algorithm follows a standard routine with optional extensions and additional features. We provide a number of methods for sampling from the prior subject to a likelihood constraint.
The software SYSBIONS is available from http://www.theosysbio.bio.ic.ac.uk/resources/sysbions/
模型选择是系统生物学科学过程的一个基本部分。给定一组相互竞争的假设，我们通常希望选择最能解释观测数据的那个假设。在贝叶斯框架中，通过贝叶斯因子（证据的比率）来比较模型，其中模型的证据是数据给予该模型的支持。一个相关的兴趣点是推断定义模型的参数的分布。嵌套采样是一种用于计算模型证据以及从后验参数分布生成样本的方法。
我们展示了一种基于C语言、GPU加速的嵌套采样实现，该实现专为生物学应用而设计。该算法遵循具有可选扩展和附加功能的标准例程。我们提供了许多在似然约束下从先验中采样的方法。
软件SYSBIONS可从http://www.theosysbio.bio.ic.ac.uk/resources/sysbions/获取。
