Semantic annotation of literature using ORKG
Purpose:
- Add your own papers and annotate them in a machine-interpretable way to increase findability and re-use of your work
- Use the ORKG templates to systematically annotate papers for meta-analysis
- Create on-demand interactive syntheses of the literature
About
Authors are the best people to tell us what their study and data are about, and they do it very well for humans reading their papers. However, there is no standardised approach to summarise essential characteristics of papers in a way that is machine interpretable and interoperable. Huge efforts have been made to standardise metadata for datasets (e.g. EML), or to standardise bibliometric information, but nothing for annotating the scientific scope and knowledge contained in publications themselves.
The ORKG makes scientific knowledge human- and machine-actionable and thus enables completely new ways of machine assistance. With the ORKG, authors, but also readers assembling papers for a meta-analysis for instance, can annotate the scientific scope and content of studies. This will help researchers find relevant contributions to their field and create state-of-the-art comparisons and reviews.
ORKG templates for Ecology and Invasion Biology
In collaboration with the ORKG, we established a set of semantic templates to guide ecologists in annotating papers in a machine-interpretable and standardised way. We built these templates in an iterative process in a series of Hi Knowledge workshops.
Semantic templates promise:
- Guidance for ecologists to choose which concepts are important to annotate
- Standardisation of annotations (allowing for controlled vocabulary, handling of synonymy, homonymy, different languages, etc.)
- Flexibility by handling heterogeneity in entries (e.g. different taxonomic levels, different units)
- Machine-interpretable and open metadata to replace publisher keywords
- Re-usability of annotations in future meta-analysis
These templates range from a very broad ecological scoping of a study (#1), similar to what is usually included in a publication's keywords, to more detailed description of study systems (#2) and study designs (#3).
We also designed templates specific to invasion biology studies. The first (#4) is a general description of the main theme, research questions, hypotheses and invasive taxa investigated, following our current conceptual scheme for invasion biology. The second (#3) describes the testing of major hypotheses in the field (described by template #6). It provides information about support or rebuttal of those hypotheses, in the same way as the Hi Knowledge data provided.
Table 1: ORKG templates created for ecology studies, with #4-6 being specific to invasion biology.
# |
Template name |
Domain |
Description |
ORKG ID |
1 |
Study in Ecology and Evolution (main template) |
Ecology |
General template for any study in the field of ecology (sensu largo). Includes: research field, approach, taxonomic coverage, geographic coverage, temporal coverage, habitat, etc. |
R593657 |
2 |
Ecological study system description |
Ecology |
Describes the properties of a specific ecological study system, which can be shared by multiple studies. |
R593670 |
3 |
Ecological study design description |
Ecology |
Describes the study design (sample size, treatment, etc.) in an ecological study. |
R593806 |
4 |
Invasion biology study research question |
Invasion biology |
Classifies the study by theme, research question, hypotheses and invasive taxa, following the scheme by Musseau et al. (in prep) |
R593830 |
5 |
Hypothesis test in invasion biology |
Invasion biology |
Annotate whether the study supports or not a major hypothesis |
R646660 |
6 |
Hypotheses in invasion biology template |
Invasion biology |
Template for describing major theoretical hypotheses in invasion biology |
R602693 |
Anyone can create a template in the ORKG!
If you think these templates are insufficient and want to create a template for your field or a particular research question, get started here.
Example application: On-demand interactive synthesis for invasion biology
Annotations in the ORKG can be extracted and used to build interactive syntheses and visualisations on a topic of interest. The easiest way to do this is to first build a comparison table across papers, choose which properties should be compared, and extract this structured data to be ingested in an R shiny app or Jupyter notebook.
As a proof-of-concept, we created an R shiny app from ORKG annotations for the Hi Knowledge dataset on 10 major hypotheses in invasion biology (Jeschke & Heger, 2018). Explore the web app here:
Hypothesis evidence explorer