A reappraisal of how to build modular, reusable models of biological systems

[...]biological information from a new experiment may raise the importance of a variable hidden within a component. [...]it becomes difficult to anticipate what information in a component should be exposed or hidden by an interface.\n For example, while one annotator may refer to the concept "D-glucose" from the ChEBI knowledge base [19], another may refer to the term "Glucose" from the Systematized Nomenclature of Medicine--Clinical Terms resource (SNOMED-CT [20]). [...]critical challenges for SAIM modeling include incentivizing model curators to annotate at this level of detail and developing intuitive software that minimizes annotation costs.