Cellular Open Resource (COR): an environment for the modelling of cardiac cellular and multi-cellular electrophysiology

心血管疾病是英国的头号杀手，每年有近238,000人死亡，占所有死亡人数的39%（国家统计局，2002年）。心血管研究的一个重要领域是心脏电生理学，研究心脏的电功能和功能障碍。大约45年前，丹尼斯·诺布尔发表了第一个心肌细胞电活动的数学模型。他随后的建模工作开发了对数据解释、假设形成、实验计划、教学有用的工具，甚至在最近的听证会上被美国食品和药物管理局（fda）接受。这种模型的进一步发展将由专门的建模环境来促进，到目前为止，这些环境的目标不是建模者就是教师，而不是大多数实验家。因此，它们要么提供强大的数值技术（主要对建模者有用），要么提供用户友好的界面（主要对教师有用），但不能同时提供这两种技术，从而降低了适用性和使用率。模型的生命周期包括几个阶段（在“黑板”上工作、实现、提交发表、实际发表和其他人使用），当涉及到模型背后的数学编码/写作时，所有这些都可能受到人为错误的影响。这显然是一个主要的问题，因为如果不是几个月，甚至是几个星期，都可能花费在追踪问题上。因此，指定了一种专用语言CellML（参见http://www.cellml.org/），现在有数百种模型可以使用这种格式。遗憾的是，只有少数环境可以执行CellML模型，允许编辑的环境就更少了。第一个支持CellML的环境，Cellular Open Resource (COR；参见http://cor.physiol.ox.ac.uk/)，已经由我们的团队开发出来。它也是第一个允许编辑和执行CellML文件的环境。从一开始，它就针对研究人员（无论是建模人员还是实验人员）和教师。今年早些时候推出了测试版，目前已经在全球33个国家使用。作为概念验证，它的重点是支持CellML（即细胞建模），并评估这样的环境是否与上述专业群体相关。COR的高录取率超出了所有人的预期。任何CellML模型的目的都是为了解决特定的问题而执行。这可能涉及在交互更改单元模型的某些属性后暂停和恢复模拟。这已经可以在COR中完成，但应该通过实现COR项目的概念来改进，这可以被视为一种描述虚拟实验的正式方式。这样的功能对建模者来说很有用，但对实验家来说也很有用，他们可以在进行相应的“湿”实验之前用它来交互式地测试一系列假设，而教师可以用它来创建教程。在后一种情况下，我们将提供一套基于牛津大学心脏细胞电生理学实践课程的教程。这不仅对该领域的教师有用，而且还说明了COR概念的实用性。涉及多个虚拟细胞的模拟（即多细胞建模）也是非常可取的（心律失常是多细胞现象）。因此，我们打算提供这样一个功能，专注于心脏电生理学，这方面的科学已经很好地建立了，这将使用最好的数值技术来实现。COR项目的概念将得到加强，以解释这种模拟的多细胞性质，并通过基于我们自己的建模工作和多细胞心脏电生理建模的基本原则的例子来说明。

Cardiovascular disease is the number one killer in the UK, with just under 238,000 fatalities p.a., and accounts for 39% of all deaths (Office for National Statistics, 2002). An important area of cardiovascular research is cardiac electrophysiology, the study of electrical function, and malfunction, of the heart. Some 45 years ago, Denis Noble published the first mathematical model of the electrical activity of a cardiac muscle cell. His subsequent modelling work has developed tools that are useful for data interpretation, hypothesis formation, experimental planning, teaching, and it was even accepted by the US Food and Drug Administration in recent hearings. The further development of such models would be facilitated by dedicated modelling environments, which have thus far been targeted at either modellers or teachers, leaving aside most experimentalists. As a result, they either offer powerful numerical techniques (mainly useful to modellers) or a user-friendly interface (mainly useful to teachers) / but not both at the same time, reducing applicability and take-up. The life cycle of a model involves several stages (work on the 'blackboard', implementation, submission for publication, actual publication and use by others), all of which being subject to human error when it comes to the coding/writing of the mathematics behind the model. This is obviously a major concern, since weeks, if not months, can literally be spent trying to track down the problem(s). A dedicated language, CellML (see http://www.cellml.org/), was therefore specified, and there are now hundreds of models that are available in that format. Sadly, there are just a handful of environments that can execute CellML models, and even fewer that allow for their editing. The first CellML capable environment, Cellular Open Resource (COR; see http://cor.physiol.ox.ac.uk/), has been developed by our group. It is also the first environment to allow for both the editing and execution of CellML files. It has, from its inception, been targeted at both researchers (be they modellers or experimentalists) and teachers. A beta version was offered earlier this year and is already being used in 33 countries worldwide. As a proof of concept, its focus has been on supporting CellML (i.e. cellular modelling) and assessing whether such an environment would be of relevance to the aforementioned groups of professionals. The high acceptance rate of COR has gone beyond all expectations. The purpose of any CellML model is to be executed to address a particular problem. This may involve pausing and resuming a simulation after having interactively changed some of the cell model's properties. This can already be done in COR, but should be improved by implementing a concept of COR Project, which could be seen as a formal way of describing a virtual experiment through time. Such a feature would be of obvious use to modellers, but also to experimentalists who could use it to interactively test a series of hypothesis before carrying out the corresponding 'wet' experiments, while teachers could use it to create tutorials. We will, in the latter context, provide a set of tutorials that are based on a practical class on cardiac cellular electrophysiology at Oxford. This will not only be useful to teachers in the field, but also to illustrate the utility of COR's concepts. Simulations that involve several virtual cells (i.e. multi-cellular modelling) are also highly desirable (arrhythmias are multi-cellular phenomena). We therefore intend to offer such a feature, focused on cardiac electrophysiology, for which the science is well established, and this will be implemented using the best numerical techniques available. The concept of COR Project will be both enhanced to account for the multi-cellular nature of such simulations and illustrated through examples based on our own modelling work and on basic principles of multi-cellular cardiac electrophysiological modelling.