Investigations
In the context of this ontology, investigations are understood to be research processes that are carried out to obtain a clearly defined output. Examples for investigations would be assessments of sex, age at death, stature, the presence or absence of specific pathological conditions, facial asymmetry or summarising parameters like prevalences of certain conditions, demographic profiles or measures of morphological variability. Research endeavours without specific expectations in terms of output would be understood as 'research projects'. Investigations follow a specific research strategy and methodology that is to be described with entities from the ontology. The purpose of these specifications is to reuse them in several investigations on different bodies of material. These investigations produce research data that are stored according to structures laid out by means of ontology elements and form a coherent body of data that can be explored with database queries. All these specifications are provided as ontology extensions that are formulated from elements from the core ontology (cf. the page on the general structure of the ontology).
The basis for modelling investigations in the RDFBones ontology is the Ontology for Biomedical Investigations (OBI). We adopt their general concepts and accommodate them to work with skeletal material from research collections.
The OBI understands investigations as processes and the class representing investigations (obo:OBI_0000066 'Investigation') is a subclass of obo:OBI_0000011 ('Planned process'). This process is divided into four sub-processes, describing subsequent stages in conducting an investigation:
Figure: Top-level processes in the representation of investigations (cf. key to symbols in network graphs).
This process describes preparations that have to be carried out before actual analyses on the material can begin. These include the definition of hypotheses, the compilation of a study design and the formulation of a plan how the study design can be applied to the material and circumstances of a specific investigation.
-> read more about investigation planning
This process comprises all interactions with the material and the generation of research data. The material is assessed for its usability in a specific investigation and specimen are selected. Observations are made, described and coded, measurements are conducted. The resulting data may be transformed or synthesised according to predefined algorithms.
-> read more about study design execution
While the execution of the study design produces research data according to set methodologies and well-defined procedures, conclusions require expert judgement. Conclusions explain what research data mean (e.g.: Was a certain trace caused by trauma or not? Did women live longer than men? What is the likelihood that an individual skeleton had a certain geographical origin?). In the context of investigations, conclusions are drawn according to predefined criteria which are documented in the study design. Potentially, different sets of criteria can be defined that would lead to diverging conclusions, even if they were based on identical data input.
This process describes how parts of the investigation output (data and conclusions) go into various publications and other documents emanating from the investigation.
The sub-processes described above represent chronological stages in conducting investigations. A study is usually planned while drafting a funding application. Once funding is secured, analyses are carried out, conclusions are drawn from the data and the results are published. It is important to note that the classes representing these subsequent processes are used rather differently when creating a record of procedures connected with a specific investigation.
The classes obo:OBI_0000039 ('Planning') and obo:OBI_0000471 ('Study design execution') most likely have one individual each associated with a specific individual of class obo:OBI_0000066 ('Investigation'). Their primary function is to bundle further sub-processes and other elements that are relevant to the planning process or the execution of a specific investigation.
Quite different from this purpose is the role that instances of the classes representing the other two principal sub-processes, obo:OBI_0000338 ('Drawing a conclusion based on data') and obo:IAO_0000572 ('Documenting'), play. Multiple conclusions can be drawn from a body of data. Consequently, class obo:OBI_0000338 ('Drawing a conclusion based on data') can have various subclasses, specifying their criteria. The OBI itself contains three such subclasses: obo:OBI_0001575 ('Assigning gene property based on phenotypic assessment'), obo:OBI_0001546 ('Comparative phenotypic assessment') and obo:OBI_0000343 ('Inductive reasoning'). Other and more specific subclasses can be defined to serve specific study designs. The instances of these classes combine a specific selection of data from study design execution with a specific conclusion, i.e. there must be one instance for each conclusion.
Accordingly, instances of obo:IAO_0000572 ('Documenting') combine a specific set of data and/or conclusions with one specific document. Each documentation process requires its own instance of obo:IAO_0000572 ('Documenting'). These processes can be classified by subclasses, the OBI providing the following: obo:OBI_0000227 ('Data encoding') and obo:OBI_0302914 ('Digital curation').
Figure: Definition of investigation types through software extensions and creation of instance data (cf. key to symbols in network graphs).
The core ontology itself does not contain any elements describing a specific type of investigation. They are exclusively implemented as ontology extensions. These extensions introduce a new subclass of obo:OBI_0000066 ('Investigation') and corresponding subclasses of the classes obo:OBI_0000039 ('Planning') and obo:OBI_0000471 ('Study design execution'). The specific subclasses for planning and study design execution are needed to link the newly defined type of investigation to a set of specifications relevant to its planning and to a set of data items relating to specimen selection and analysis.
Extensions are also likely to introduce a whole set of subclasses of obo:OBI_0000338 ('Drawing a conclusion based on data') to define a number of conclusions that are to be drawn from the collected data. In terms of obo:IAO_0000572 ('Documenting'), extensions might well do with the standard options provided by the core ontology and not necessarily define specific subclasses, though this is possible.
When a user initiates a study as an instance of the new subclass of obo:OBI_0000066 ('Investigation'), this immediately necessitates the creation of corresponding instances of the related classes representing planning and study design execution for this type of study. Consequently, these instances can be created automatically by a software application. Instances of subclasses relating to conclusions and documentation, on the other hand, define specific relations between data and documents, which need to be set manually.
