SIGNAL: Stochastic process algebra for biochemical signalling pathway analysis

计算系统生物学利用计算机对生物体进行建模来帮助我们理解那些我们无法直接看到或测量的内部工作过程。基于实验确定的数据，系统生物学家发明了一个他们认为事物如何运作的模型。通常用两种计算机模拟中的一种来分析模型：随机模拟或确定性模拟。在随机模拟中，概率论的数学理论用来表示反应发生的速度或反应物的数量的不确定程度。在确定性模拟中，常微分方程理论用于给出大量分子元素的有效连续逼近。计算模型在这里是有帮助的，因为基于实验室的实验是一个极其昂贵、耗时和劳动密集型的过程。信号转导是计算建模的一个重要课题。信号转导途径是允许细胞感知刺激并将信号传递给细胞核的生化途径，然后细胞核做出适当的反应。它们是具有内置反馈机制的复杂信号传递过程。信号通路嵌入在更大的网络中，并参与重要的过程，如增殖、细胞生长、运动、细胞通讯和细胞程序性死亡（凋亡）。功能障碍会导致许多疾病，包括癌症、糖尿病和许多其他疾病。尽管近年来在实验方面取得了巨大的进展，但仍然缺乏良好的、可预测的途径模型来指导实验和药物开发。迄今为止，模型要么编码方面，如哪些蛋白质具有相互作用的潜力，要么使用常微分方程提供系统动力学的模拟。我们将根据我们对并发计算系统建模的经验开发一种新的分析路径建模方法。关键思想是路径具有随机的、可计算的内容。我们将使用随机过程代数来建模路径，这些代数表示连续时间马尔可夫链，从而提供新的定量分析和新的方法来构建路径和不完全行为的原因。

The science of computational Systems Biology uses computer modellingof living organisms to help us understand those process at work insidewhich we cannot directly see or measure. Based on experimentallydetermined data, a systems biologist invents a model of how they thinkthat things work. The model is usually analysed by one of two kindsof computer simulation: stochastic or deterministic. In a stochasticsimulation the mathematical theory of probability is used to express adegree of uncertainty about how fast reactions happen or thequantities of reactants which are in use. In a deterministicsimulation the theory of ordinary differential equations is used togive an efficient continuous approximation of very large numbers ofmolecular elements. Computational modelling is helpful here becauselaboratory-based experimentation is an extremely expensive,time-consuming and labour-intensive process. An important subject forcomputational modelling is signal transduction.Signal transduction pathways are biochemical pathways which allowcells to sense a stimulus and communicate a signal to the nucleus,which then makes a suitable response. They are complicated signallingprocesses with built-in feedback mechanisms. Signalling pathways areembedded in larger networks and are involved in important processessuch as proliferation, cell growth, movement, cell communication, andprogrammed cell death (apoptosis). Malfunction results in a largenumber of diseases including cancer, diabetes and many others. Despiteenormous experimental advances in recent years there is still anabsence of good, predictive pathway models which can guideexperimentation and drug development. To date, models either encodestatic aspects such as which proteins have the potential to interact,or provide simulations of system dynamics using ordinary differentialequations.We will develop a novel approach to analytic pathway modelling basedon our experience of modelling concurrent computing systems. The keyidea is that pathways have stochastic, computational content. We willmodel pathways using stochastic process algebras which denotecontinuous time Markov chains thus affording new quantitative analysisand new ways to structure pathways and reason about incompletebehaviour.