- About
- What are ontologies and how do they improve data quality?
- The Contextual Data Specification Package
- Contacts
- License
- Acknowledgements
As antimicrobial resistance (AMR) becomes more widespread, efforts to extend the usability of current antimicrobial agents by developing and applying diagnostics to identify the appropriate narrow-spectrum antimicrobial for each infection take centre stage. However, diagnostic developers currently suffer from late-stage attrition, creating a “valley of death” that prevents diagnostic tools from reaching patients. The B2B2B AMRDx network is a cross-disciplinary, geographically diverse and gender-balanced group of partners from universities (bench), hospitals (bedside), for-profits (business), governments and nonprofits (beyond), with expertise in all One Health settings: human, animal, and environmental AMR. Improved communications between stakeholders in multiple sectors, an objective of the WHO’s Oslo Medicines Initiative, benefits the development of rapid, reliable AMR diagnostics.
The B2B2B project focuses on 3 objectives to address the challenges faced by AMR diagnostics developers.
- Objective 1 expands an open, self-curated, comprehensive online AMR Diagnostics Developer Directory (ADDD) to facilitate exchange of ideas and create synergies in the field.
- Objective 2 further develops the JPIAMR Seq4AMR Virtual Benchmarking Platform (VBP) for genotype-to-phenotype microbial benchmarking studies that will include gold standard whole genome sequences, phenotypes, and contextual data (metadata), as well as accompanying microbial isolates.
- Objective 3 defines policy Pathways to AMR-specific Incentives for Developing Diagnostics (PAIDD), maximising public health benefits from the use of AMR diagnostics.
As part of Objective 2 of the B2B2B project - to provide gold standard contextual data to accompany benchmark datasets - the B2B2B data specification was developed. The B2B2B data standard is an ontology-based One Health AMR data standard for bacterial pathogens, that provides standardized fields, pick lists of controlled vocabulary, and prescribed formats for the harmonized capture of contextual data. The B2B2B vocabulary was sourced from the GRDI-AMR-One-Health data standard. The standardized fields are based on community standards such as NCBI’s combined Pathogen and Environmental attribute package derived from internationally agreed upon Minimal Data for Matching (MDM) standards, as well as applicable fields from different MIxS packages (Genomic Standards Consortium).The specification is also ISO 23418 compatible (Microbiology of the food chain — Whole genome sequencing for typing and genomic characterization of bacteria — General requirements and guidance).
The goal of the specification is to enable the creation of richly annotated datasets that can be applied to a wide variety of AMR prevention and innovation use cases. It should be noted that new terms can be added to the standard by submitting a GitHub issue request (see New Term Request SOP below), and that additional standardized attributes are available (see the GRDI-AMR repository for more information).
What are ontologies and how do they improve the quality of contextual data in gold standard benchmark AMR datasets?
Labs collect, encode and store information in different ways. They use different fields, terms and formats, they categorize variables in different ways, and the meanings of words change depending on the focus of the organization (think of the word “plant”. To someone in agriculture, “plant” could mean an organism that carries out photosynthesis, while a food regulator might understand the word “plant” to mean a factory where food products are made). This variability makes comparing, integrating and analyzing data generated by different organizations like trying to compare apples, oranges and bananas, which is difficult to do.
Ontologies are collections of controlled vocabulary that are arranged in a hierarchy, where all the terms are linked using logical relationships. Ontologies are open source and meant to represent “universal truth” as much as possible (so not tied to one organization’s vocabulary of use case). Ontologies encode synonyms, which enables mapping between the specific languages used by different organizations, and every term in the ontology is assigned a globally unique and persistent identifier. Using ontology terms to standardize B2B2B2 contextual data not only helps make data more interoperable by using a common language, it also helps to make contextual data FAIR (Findable, Accessible, Interoperable, Reusable).
This specification is implemented via a spreadsheet-based data collection instrument (i.e. metadata template) in the validation tool the DataHarmonizer, with accompanying Field and Term reference guides (which provide definitions and additional specific guidance). New terms and/or term changes can be requested using issue request forms, with additional guidance on how to do so outline in the New Term Request (NTR) SOP (please note that the specification will be updated periodically to address user need). This resources are available in the files of this repository and listed below under Package Contents.
Please note that development of the specification is dynamic and it will be updated periodically to address user needs. Versioning is done in the format of x.y.z.
x = Field level changes
y = Term value / ID level changes
z = Definition, guidance, example, formatting, or other uncategorized changes
Descriptions of changes are provided in [release notes](https://github.com/cidgoh//releases) for every new version.
- Pathogen Genomics Package (B2B2B2)
- Template schema files can be found as
.yaml/.json/.tsvunder pathogen-genomics-package/templates/b2b2b2
- Template schema files can be found as
- DataHarmonizer App
- The DataHarmonizer is a standardized browser-based spreadsheet editor and validator.
- Instructions on "Getting Started" downloading and using the application can be found under DataHarmonizer Instructions and SOP below.
- Further information about application functionality can be found on the DataHarmonizer Wiki.
For more information and/or assistance, contact at or submit a repository issue request.
Pending / To Be Determined
Brought to you by The Centre for Infectious disease Genomics and One Health