FuTRES is a semantic database for the storage and retrieval of vertebrate trait data. Using community-developed ontologies, it links measurable trait data, such as body mass or humeral length, to a single specimen and encodes them into RDF triples for easy searchability and accessibility. RDF triples link instances of ontology classes to each other as well as to literals (values or strings), thereby creating a graph which captures data which is computable using automated reasoners. FuTRES does not have field limitations the way a relational database does. Any new trait can be added to FuTRES and put in the context of existing anatomical, geographic, temporal, and taxonomic knowledge in the database.
We have created a pipeline for extracting trait data from the VertNet database as well as dark data from publications and from unpublished data stores (including the Bernor equid database). We have assembled a database of over 1 million specimen records and ~150 traits from paleontological, zooarchaeological, and modern biological sources. We have initially limited the scope to mammals to test our workflows for data ingestion and retrieval. The database will become available through both a searchable portal and an API by the end of 2021. We will build an R package to facilitate data access from the API.
As an example of potential uses of FuTRES, we have created new allometric scaling regressions from individual body mass and body length measurements from > 98,000 museum records of modern mammals. In this way, we investigate the extent to which large samples would allow more precise estimation of body mass of extinct mammals through similar allometric scaling functions.
Results/Conclusions
We found a significant relationship between log mass and log length, with p < 10-15 and r2 = 0.96. Body length, available from VertNet, stands in for the dental or skeletal characters that would be used in a paleobiological analysis. As researchers contribute data to FuTRES, it will become possible to develop similar regressions using skeletal and dental measurements.
Over time, we will extend the scope of FuTRES to more diverse taxa and data types, expanding support to all vertebrates, to trait types inferred from images (e.g., mesowear, microwear, and 3D morphometric data), and to individual-based interaction events. We strive to create a community-driven resource that allows biologists to share and discover trait data as easily as we can now share and discover genetic data.