Closed javadch closed 1 year ago
Data Analysis and Table Priority To build an ontology, we first started with the data analysis. Data analysis means what kind of elements are there in the relational database. All total, we have 96 tables. We made a priority list based on the importance of the conversion of the table to RDF. Overall, there is a total of 2294578 records in the database. The SQL dump is available at https://nextcloud.dsmz.de/s/dwgPo2FskjCAMNX?path=%2FBacDive%20Dump
We decided to work on the table having a priority list higher than 0, so at the end, we need to convert 71 tables.
The priority list is available at https://github.com/TIBHannover/diaspora/tree/main/wp2/t2.1
A glimpse of Schema of BacDive database
Ontology Development
We used the YAMO methodology by following the MOD guidelines to design the ontology. We aligned the ontology by following an upper ontology in order to make the ontology interoperable. It clearly identifies the semantics for very common terms, which play a major role in the vocabulary used in the domain discipline. DOLCE(a descriptive ontology for linguistic and cognitive engineering)} is used as the upper ontology for alignment.
The ontology file is available at https://github.com/TIBHannover/diaspora/tree/main/wp2/t2.1
Mapping Rules
For data mapping, we have used the SDM RDFIzer(https://github.com/SDM-TIB/SDM-RDFizer), which creates a mapping rule for converting tabular data into triples, and the detailed instructions are given below.
An example mapping rules is available at https://github.com/TIBHannover/diaspora/tree/main/wp2/t2.1
Ontology publication
• TIB would be ok, but maybe only as a second place, as it has not enough reach so far • I would favour BioPortal atm, which is handy and to my perception widely known. • Obo foundry is also well-established -> do we have overlap with existing ontologies? I would guess so? • OLS at EBI is smaller, still interesting.
BD_table | Priority (10 is high) | short description | Foreign Keys
To deal with this, we followed an Agile approach in which we started with a small prototype then we plan to scale it for all BacDive database. In the following section, I will explain the steps used for the prototype. Currently, we are in the process of scaling our approach.
BD_table | Priority (10 is high) | short description | Foreign Keys