Questions and Answers: Ontology-based species descriptions

What is an ontology?

Ontologies are computer technologies, such as databases or computer programs, that are used to represent and reason about knowledge in a particular domain. It could be, for example, knowledge of the anatomy and morphology of some group of organisms. The knowledge represented in an ontology can be interpreted and understood by computers. Roughly speaking, knowledge in ontologies is represented by means of networks in which nodes represent concepts and edges the relationships between those concepts. Hence, ontologies can be seen as technologies for organizing and structuring knowledge.

However, ontologies are not merely conceptual networks. They are also equipped with a distinctive feature that makes them powerful tools for knowledge inference. This feature is the mathematical logic that provides a formal framework for expressing and reasoning about concepts and relationships in ontologies. The knowledge encoded in an ontology is structured and formalized in a way that conforms to the precise and well-defined language of mathematical logic. This ensures that the ontology can be reasoned about in a rigorous and systematic manner, allowing for reliable inference of knowledge not explicitly stated in the ontology or data.

Is there a connection between the term ontology in computer science and philosophy?

Yes, the term ontology also refers to a branch of philosophy that deals with the study of existence, reality, and relationships between entities. Due to their conceptual similarities, computer science borrowed the term 'ontology' from philosophy.

When it comes to ontological technologies, why is the term "semantic" frequently used?

The term "semantic" is frequently used in conjunction with ontologies since ontologies represent the logical meaning of concepts. In other words, ontologies are concerned with the semantics of concepts within a particular domain, which can be understood by computers. As an example, a semantic species description is one that may be automatically analyzed by a computer.

What are the differences between ontologies and knowledge graphs?

Both terms are related. An ontology specifies the concepts and relationships that are general and constant within a particular domain, and they usually apply to all instances of data in that domain. On the other hand, a knowledge graph represents specific instances of data using concepts encoded in an ontology or set of ontologies. Therefore, knowledge inference provided by ontologies can also be applied to knowledge graphs representing specific data. As an example, an insect anatomy ontology may include the statement "all insects have six legs", whereas knowledge graphs may refer to the legs in a particular species, for instance, "legs in Tribolium castaneum are brown".

How can ontologies be used to study phenotypic data and species descriptions?

Ontologies can be used to represent general knowledge about organismal phenotypes. Those ontological concepts can be used to develop knowledge graphs that describe species and their phenotypes. In this graph, nodes represent phenotypic characteristics and edges represent their relationships. A morphologist would define a node in such a graph as a particular anatomical observation about an organism or group of organisms. For example, the observation that a cat has claws is represented by a node in a knowledge graph. Furthermore, nodes can store (meta)data about morphological (or other) observations, such as length of morphological structure, voucher id, or any other type of information. Edges are used to indicate how nodes are related to one another. For instance, cats have claws located on their toes and these claws have the characteristic of being sharp. 'Located on' and 'have characteristics' are the edges in this case. Phenotypes are typically described by using a dozen different ontologies. Thus, semantic phenotypes can be understood by computers through the use of ontological technologies.