Skip to content

Latest commit

 

History

History
276 lines (164 loc) · 21 KB

README.md

File metadata and controls

276 lines (164 loc) · 21 KB
Raw

The COVID-19 epidemiology and monitoring ontology

License: CC0-1.0

The COVID-19 Epidemiology and Monitoring Ontology (CEMO) is designed to make epidemiological quantitative data for monitoring the COVID-19 outbreak machine-readable and interoperable to facilitate its exchange, integration and analysis, to eventually support evidence-based rapid response. This ontology has built following knowledge-engineering standards and the OBO principles to bridge epidemiology into the semantic landscape of the biomedical sciences. See a detailed description of the model and the axiom patterns implemented in the following section.

This is the first release of the ontology, which will be refined iteratively with domain experts and users. This project was conceived during the virtual BioHackathon COVID-19 2020 (BH-COVID19 2020 GitHub) and started at the ELIXIR-Europe BioHackathon 2020 (BH-EU 2020 GitHub), see project description below, and continued during the SWAT4HCLS virtual hackathon 2021. This project is the result of the commitment and volunteering work of scientists, experts and any person with interest in contributing.

Note that CEMO is currently under development and it is intended to be submitted to OBO at some later date, which means that the URIs may not be stable.

What is next?

This ontology is under development. Currently, we are working on:

  • Ontological model: Improving the logical structure of the model and text content.

  • OBO Foundry: Better following the OBO principles as we intend to submit the ontology to the OBO Foundry.

  • COVID-19 harmonization: We are collaborating with the COVID-19 harmonization community effort for the harmonization of epidemiological information.

  • New challenges: we are evaluating the model for new challenges.

I want to give you feedback

Please leave us a comment on the issue tracker.

I want to contribute

Feel free to fork & make PR to contribute!

Note that CEMO is currently under development and subject to change.

Please leave us a comment on the issue tracker if you want to be involved.

I want to cite

You can use the following citations:

  • SWAT4HCLS 2023 Conference: Presentations, Posters and Supplementary Materials. Poster DOI PDF

  • ISMB/ECCB 2021 Bio-Ontologies track abstract - 2021 ISMB Bio-Ontologies Community of Special Interest Zenodo Collection DOI PDF

  • BioHackrXiv DOI PDF

Formal description (under development stay tuned!)

CEMO is designed to be an OBO ontology, i.e., founded on the BFO hierarchy, and formalized through SIO and GFO upper-level ontologies. The taxonomic structure is extended from IDO, a core ontology for infectious diseases. The ontology is built in OWL 2, a DL-based formalism and semantic web standard for knowledge representation to enable data sharing and logic reasoning.

Our formal modeling followed a rationale already used in other studies: 1) determine the domain and scope of the ontology; 2) ontology reuse and addressing poor ontological coverage of COVID-19 epidemiology; and 3) development of a conceptual model.

Domain scope

We extracted core domain knowledge concepts from Epidemiology Surveillance reference books [Ferran Martínez Navarro et al. Vigilancia Epidemiológica. McGraw-Hill Interameri-cana, 2004. ISBN:84-486-0245-5][B. MacMahon and D. Trichopoulos. Marbán SL, Harvard Medical School of Public Health, Boston, Massachussetts. ISBN:84-7101-317-7] and a handbook of Epidemiology. The list of core concepts for interpretation are: person, cohort, population, time, space, infection, line of infection, disease, line of disease, transmission, exposure, risk factor, contagiousness, virulence, source of infection, distribution, case, natural history, etiology, and infectious dose.

Re-use

We re-used ontological terms and models as much as possible. Specifically we re-used:

  1. Terms from IDO for the infectious disease domain and from STATO, APOLLO_SV and GENEPIO for epidemiology. We mapped the manually extracted list of terms to these ontological terms using ontology search engines such as Ontobee and Ontology Lookup Service - EBI.
  2. Quantitative Models for quantitative data, specifically: our minimal model for clinical measurements presented at the ISMB/ECCB Bio-Ontologies COSI 2020; the SIO design pattern for measurements; the EJP RD Core model for patient registries data; and the Beat-COVID model for health data in the LUMC.
  3. Mortality Model based on the GFO foundational ontology and Morbidity and Mortality data in databases for epidemiology research in Brazil Santana F, Freitas F, Fernandes R, Medeiros Z, Schober D. Towards an ontological representation of morbidity and mortality in Description Logics. J Biomed Semantics. 2012;3 Suppl 2(Suppl 2):S7. doi:10.1186/2041-1480-3-S2-S7.
  4. Person-Patient Relationship Description for biomedical research, specifically the GA4GH Phenopackets standard and the OMOP common data model.

Ontological model

We developed a conceptual model for quantitative epidemiological data. This model is based on the minimal model for clinical measurements; the SIO Design Pattern for measurement; the EJP RD CDE Core model; and the Beat-COVID model for health data in the LUMC.

In an outbreak, descriptive epidemiology or the calculation of different measures of frequency that depend on person, time and space parameters, where entities under study are populations or patients, is first applied to characterize and monitor it. A graphical representation of the model:

Use case: analytical epidemiology

In analytical epidemiology are frequent observational studies to establish causal-effect relations, where the disease-risk factors association is evaluated. A graphical representation of the model extended for this use:

Axiom patterns

We implemented axiom patterns to describe epidemiological quantitative parameters and timelines domain concepts. For quantitative parameters we added the following axiomatization (in OWL Manchester Syntax):

"epidemiology quantitative quality" SubClassOf (*is_attribute_of* some "entity")

See what entity mean:

* `Entity`: we use one of the GFO-based concepts to link to epidemiological parameters for descritive epidemiology: *person*, *time*, *space*, *line of infection* or *line of disease*.

For instance:

"incidence" SubClassOf (*is_attribute_of* some "space")

For timelines we re-used GFO using the 'chronoid' concept and the GFO-based 'mortality' model approach with the following axiomatization:

"incubation period" SubClassOf "process"
                    and (*caused_by* only "infectious process")
                    and (*has_participant* only "person")
                    and (*projects_to* exactly 1 "chronoid")

For this pattern we needed to integrate the time concept as an individual item and its subclass chronoid in the ontology. Process concept in BFO is logically equivalent to Process concept in GFO. This is a first approach, we aim to extend it in further iterations of the model to chain temporal periods in an ordered and formal description of the timeline.

Formal description of patient-population link

Following RDA best practices recommendation to link patient-population information, we re-used the relationship in the GA4GH Phenopackets standard based on composition semantics and formalized it with the axiom:

"population" *has_member* "person"

To allow reasoning for discovery in translational research, we aim to extend this description to more expressive DL-based definition.

FAIR ontology

We base our decisions on building this ontology to follow as much as possible the FAIR principles.

Summary

Projects using CEMO

The CEMO ontology is currently being used in:

  • FAIRification of Leiden Longevity Study (LLS): The aim of the study was to retrieve genetically determined mechanisms of longevity and their interaction with the environment. CEMO is used for the FAIRification of the collected epidemiological data. Especifically, data is semantically modelled based on the CEMO model, which is the OBO version of the EJP RD core model, i.e., it is the EJP RD core model mapped to ontological terms from ontologies currently accepted in the OBO Foundry. To know more about the LLS-FAIRification project see the GitHub repository.

Authors

Núria Queralt-Rosinach [ [email protected] ], Paul N. Schofield, Robert Hoehndorf, Claus Weiland, Marco Roos

License

CEMO is licensed under the terms of the CC0 License.

Acknowledgements

Special thanks to Elena de la Calle, Philippe Rocca-Serra, Rajini Nagrani (RDA), Niek van Ulzen (LLS), Charles Tapley Hoyt (OBO) for their input. BioHackathon COVID-19 2020, SWAT4HCLS 2021, and BioHackathon-Europe 2021 for supporting this project. This initiative has received funding from the EU's Horizon 2020 research and innovation programme under grant agreement No 825575 (EJP RD), the collaboration project Trusted World of Corona (TWOC) co-funded by the PPP allowance by Health-Hollland, and the Alan Turing Institute.

INCEPTION

Epidemiology and monitoring ontology for COVID-19 project - Virtual BioHackathon 2020 project

Abstract

Epidemiological data is necessary to monitor public health, and to assess the impact of disease outbreaks and efficacy of mitigating interventions. In the context of an infectious disease outbreak it is imperative to have these data as FAIR as possible to facilitate rapid analysis to support timely evidence-based decision making. During the past virtual BioHackathon-COVID-19 we evaluated the availability of some epidemiological quantitative indicators in OBO ontologies and detected that while between the Statistics ontology and Epidemiology Ontology (EPO) we have many of the usual concepts described, essential classes are entirely missing for capturing these different indicators with the precision required, and also definition reconciliation needs to be done. Furthermore, we noticed that EPO is not maintained since its publication and has been deprecated from OBO Foundry [1]. Another issue that may be improved is the current absence of axioms and definition patterns that relate epidemiology (i.e., observations of a population) to clinical ontologies (i.e., observations on an individual). A formal model is needed to describe epidemiological data for monitoring events such as COVID-19 disease as well as to capture information related to other disease outbreaks and future epidemics.

References

[1] Pesquita et al. The epidemiology ontology: an ontology for the semantic annotation of epidemiological resources, J Biomed Semantics. 2014; 5: 4. doi: 10.1186/2041-1480-5-4

Goals

Our main goal is to provide a minimal epidemiological ontology developed during the hackathon days that would have application to the entire biomedical community, not only the life science communities. The goals of the proposed project are:

  • To identify and structure the missing concepts, extending current ontologies or making a unified model combining them into single ontology.

  • To enable the biomedical community to provide machine-readable quantitative epidemiological data and make it easier for scientists to find and share it, and to link patient data to infectivity, transmissibility and outcomes of SARS-CoV-2 infection in humans.

  • To ultimately facilitate surveillance, research, and policy-making in pandemics like COVID-19.

Our project will facilitate epidemiological and bioinformatic studies enabling machine-readable and interoperable epidemiological data through common ontologies with application on patient data. Many of the solutions we expect to develop will be of general utility and we expect that the unified, or extended and axiomatised ontology will integrate as a bridging and facilitating component into the semantic landscape of the biomedical sciences.

Topics

Covid-19 | Data Platform | Federated Human Data | Interoperability Platform

Team

Lead(s)

Núria Queralt-Rosinach [ [email protected] ],

Contributors

Elena de la Calle [ [email protected] ], Philippe Rocca-Serra [ [email protected] ]

Expected outcomes

Develop an epidemiology information model in OBO focused on quantitative indicators. Incorporate this model in an existing resource for the community.

Plan

From the first list of epidemiological quantitative terms analyzed, our plan is to:

  1. add new COVID-19 relevant content through manual extraction from publications in medrXiv and or case report forms
  2. try to map these to existing OBO ontologies
  3. refine the list of terms with epidemiologists
  4. define and implement axiom patterns
  5. contact ontology stakeholders to define what is missing and either to ask for extensions to existing ontologies or to build a new, logically well-formed, and accurate ontology in OBO.

We plan to perform two rounds of work, a first round to focus on quantitative indicators and a second round to expand to other indicators such as WHO indicators.

Current outcomes

  1. list of relevant terms manualy curated from publications, case report forms and institutional official websites [collaborative gsheet]
  2. formal ontology in owl under development [cemo.owl]

References for curation and related models

A Review terms

1st round: quantitative terms

Review sources (publications, medrXiv, CRFs):

2nd round: other epidemiology terms

Expand to other indicators such as WHO indicators:

B Review models - design patterns

  1. Statistics Ontology (STATO) [website] (contact Philippe Rocca-Serra)
  2. Epidemiology Ontology (EPO)
    deprecated
    [ontobee]
  3. Infectious Disease Ontology (IDO) [ontobee]
  4. CoVoc Coronavirus Vocabulary (CoVoc) [EBI OLS]
  5. Genomic Epidemiology Ontology (GenEpiO) [ontobee]
  6. Ontology for Nutritional Epidemiology [[NCBO BioPortal](Ontology for Nutritional Epidemiology)]
  7. Bringing epidemiology into the Semantic Web [NERO]
  8. Referenced by Claus Weiland: [Towards an ontological representation of morbidity and mortality in Description Logics]
  9. Expressive probabilistic description logics [DOI]

General references

  2. WHO indicators
  3. From Kees van Bochove (twitter, the hyve::ohdsi): Have a look at the way we used schema.org to express structured metadata about observational studies model and example study
  4. On the usefulness of ontologies in epidemiology research and practice [PMC ID]