本体论是什么

Dispite this popularity and widespread use, many aspects of the Gene Ontology remain poorly understood, at times even by experts. For instance, unbeknownst to most users, routine procedures such as GO term enrichment analyses remain subject to biases and simplifying assumptions that can lead to suprious ocnclusions. -----Christobe Dessimoz and Nives Skunca

一、本体论的概念

本体论是一个以结构化的可计算的格式来描述某个感兴趣领域里实体和其关系的结构（Ontologies are computatonal structures that describe the entities and relationships of a domain of interest in a structured computable format, which allows for their use in multiple applications），所以其核心在于实体（分类）的集合和相应的关系。那么基因本体论的核心就在于基因研究这个领域的核心实体和对应的关系。

二、本体论中的元素

1、classes（类别）

class是本体论的基本单元。它代表了某个特定领域里的一类东西，通常与本体论中命名空间里唯一的标识符关联。如GO:0006423。为了保证其稳定性，这类标识符不包含类别名称或定义的参考（they do not contain a reference to the class name or definition）。应该就是说GO号并没有特定的含义，只是字母和数字的组合，只有当它们与元数据关联时才有明确的含义，所以GO号永远可以不变，只需要变动元数据就可以了。

2、metadata（元数据）

元数据就是解释性的文本数据，其实就是与其关联的GO号的定义，该定义应该非常明确，而且能提供足够的信息（sufficiently distinguishing different classes in an ontology so that a user can determine which is the best to use for annotation）。每个GO号可以对于多个元数据。

3、relations（关系）

类别通常以层级的关系排列，类似与数据结构中的树，但又不是简单的树，是一种有向非循环图，该图应该避免循环。通常常用的关系类型有：

relationship type	informal meaning	examples
part_of	部分关系	a brain is part_of a body
derives_from	来源于两个完全不同的实体，有短暂的继承关系	a zygote derives_from a sperm and an ovum
has_participant	连接实体的关系	an apoptotic process has_participant a cell
has_function	连接物质实体和功能的关系	an enzyme has_function to catalyse a specific reaction type

4、formats（格式）

目前GO的储存格式是human-readable Open Biomedical Ontologies(OBO)，还有一种新发展出来的Semantic Web standard Web Ontology Language(OWL)

5、axioms（公理）

也就是内含的逻辑含义。比如羧酸是一种碳酮酸，隐含的逻辑就是任意的羧酸都是碳酮酸。OWL语言中的逻辑构架（直接从书中复制）：

language component	informal meaning	examples
quantification:universal(only) or existential(some)	universal quantification: all relationships of that type the target has to belong to the specified class; existential quantification:at least one member of the target class must participate in a relationship of that type	molecule has_part some atom
cardinality:exact, minimum or maximum	It is possible to specify the number of relationships with a given type and target that a class must participate in, or a minimum or maximum number thereof	human has_part exactly 2 leg
logical connectives:intersectin or union	It is possible to build complex expressions by joining together parts using the standard logical connectives and, or.	vitamin B equivalentTo
Negation(not)	in addition to building complex expressions using the logical connectives, it is possible to compose negations	tailless equivalentTo not
Disjointness of classes	It is possible to specify that classes should not share any members	organic disjointFrom inorganic
equivalence of classes	It is possible to specify that two classes--or class expressions-- are logically equivalent, and that they must by definition thus share all their member	melanoma equivalentTo skin cancer

三、缺点

本体论是基于逻辑的，它可以很好地表示Ture or False，但不能表示模糊的，在一定条件下的概念。同时，它也很难表示动态的信息。

参考文献

1、geneontology.org/docs/ontology-documentation/
2、The Gene Ontology Handbook