DataMed - an open source discovery index for finding biomedical datasets

Xiaoling Chen; Anupama E. Gururaj; Burak Ozyurt; Ruiling Liu; Ergin Soysal; Trevor Cohen; Firat Tiryaki; Yueling Li; Nansu Zong; Min Jiang; Deevakar Rogith; Mandana Salimi; Hyeon eui Kim; Philippe Rocca-Serra; Alejandra Gonzalez-Beltran; Claudiu Farcas; Todd Johnson; Ron Margolis; George Alter; Susanna Assunta Sansone; Ian M. Fore; Lucila Ohno-Machado; Jeffrey S. Grethe; Hua Xu

doi:10.1093/jamia/ocx121

DataMed - an open source discovery index for finding biomedical datasets

Xiaoling Chen, Anupama E. Gururaj, Burak Ozyurt, Ruiling Liu, Ergin Soysal, Trevor Cohen, Firat Tiryaki, Yueling Li, Nansu Zong, Min Jiang, Deevakar Rogith, Mandana Salimi, Hyeon eui Kim, Philippe Rocca-Serra, Alejandra Gonzalez-Beltran, Claudiu Farcas, Todd Johnson, Ron Margolis, George Alter, Susanna Assunta SansoneIan M. Fore, Lucila Ohno-Machado, Jeffrey S. Grethe, Hua Xu

Artificial Intelligence and Informatics

Research output: Contribution to journal › Article › peer-review

Abstract

Objective: Finding relevant datasets is important for promoting data reuse in the biomedical domain, but it is challenging given the volume and complexity of biomedical data. Here we describe the development of an open source biomedical data discovery system called DataMed, with the goal of promoting the building of additional data indexes in the biomedical domain. Materials and Methods: DataMed, which can efficiently index and search diverse types of biomedical datasets across repositories, is developed through the National Institutes of Health-funded biomedical and healthCAre Data Discovery Index Ecosystem (bioCADDIE) consortium. It consists of 2 main components: (1) a data ingestion pipeline that collects and transforms original metadata information to a unified metadata model, called DatA Tag Suite (DATS), and (2) a search engine that finds relevant datasets based on user-entered queries. In addition to describing its architecture and techniques, we evaluated individual components within DataMed, including the accuracy of the ingestion pipeline, the prevalence of the DATS model across repositories, and the overall performance of the dataset retrieval engine. Results and Conclusion: Our manual review shows that the ingestion pipeline could achieve an accuracy of 90% and core elements of DATS had varied frequency across repositories. On a manually curated benchmark dataset, the DataMed search engine achieved an inferred average precision of 0.2033 and a precision at 10 (P@10, the number of relevant results in the top 10 search results) of 0.6022, by implementing advanced natural language processing and terminology services. Currently, we have made the DataMed system publically available as an open source package for the biomedical community.

Original language	English (US)
Pages (from-to)	300-308
Number of pages	9
Journal	Journal of the American Medical Informatics Association
Volume	25
Issue number	3
DOIs	https://doi.org/10.1093/jamia/ocx121
State	Published - Mar 1 2018

Keywords

Data discovery index
Dataset
Information dissemination
Information storage and retrieval
Metadata

ASJC Scopus subject areas

Health Informatics

Access to Document

10.1093/jamia/ocx121

Cite this

Chen, X., Gururaj, A. E., Ozyurt, B., Liu, R., Soysal, E., Cohen, T., Tiryaki, F., Li, Y., Zong, N., Jiang, M., Rogith, D., Salimi, M., Kim, H. E., Rocca-Serra, P., Gonzalez-Beltran, A., Farcas, C., Johnson, T., Margolis, R., Alter, G., ... Xu, H. (2018). DataMed - an open source discovery index for finding biomedical datasets. Journal of the American Medical Informatics Association, 25(3), 300-308. https://doi.org/10.1093/jamia/ocx121

Chen, X, Gururaj, AE, Ozyurt, B, Liu, R, Soysal, E, Cohen, T, Tiryaki, F, Li, Y, Zong, N, Jiang, M, Rogith, D, Salimi, M, Kim, HE, Rocca-Serra, P, Gonzalez-Beltran, A, Farcas, C, Johnson, T, Margolis, R, Alter, G, Sansone, SA, Fore, IM, Ohno-Machado, L, Grethe, JS & Xu, H 2018, 'DataMed - an open source discovery index for finding biomedical datasets', Journal of the American Medical Informatics Association, vol. 25, no. 3, pp. 300-308. https://doi.org/10.1093/jamia/ocx121

@article{b8432d0a1f8f43fe892cd277a077166b,

title = "DataMed - an open source discovery index for finding biomedical datasets",

abstract = "Objective: Finding relevant datasets is important for promoting data reuse in the biomedical domain, but it is challenging given the volume and complexity of biomedical data. Here we describe the development of an open source biomedical data discovery system called DataMed, with the goal of promoting the building of additional data indexes in the biomedical domain. Materials and Methods: DataMed, which can efficiently index and search diverse types of biomedical datasets across repositories, is developed through the National Institutes of Health-funded biomedical and healthCAre Data Discovery Index Ecosystem (bioCADDIE) consortium. It consists of 2 main components: (1) a data ingestion pipeline that collects and transforms original metadata information to a unified metadata model, called DatA Tag Suite (DATS), and (2) a search engine that finds relevant datasets based on user-entered queries. In addition to describing its architecture and techniques, we evaluated individual components within DataMed, including the accuracy of the ingestion pipeline, the prevalence of the DATS model across repositories, and the overall performance of the dataset retrieval engine. Results and Conclusion: Our manual review shows that the ingestion pipeline could achieve an accuracy of 90% and core elements of DATS had varied frequency across repositories. On a manually curated benchmark dataset, the DataMed search engine achieved an inferred average precision of 0.2033 and a precision at 10 (P@10, the number of relevant results in the top 10 search results) of 0.6022, by implementing advanced natural language processing and terminology services. Currently, we have made the DataMed system publically available as an open source package for the biomedical community.",

keywords = "Data discovery index, Dataset, Information dissemination, Information storage and retrieval, Metadata",

author = "Xiaoling Chen and Gururaj, {Anupama E.} and Burak Ozyurt and Ruiling Liu and Ergin Soysal and Trevor Cohen and Firat Tiryaki and Yueling Li and Nansu Zong and Min Jiang and Deevakar Rogith and Mandana Salimi and Kim, {Hyeon eui} and Philippe Rocca-Serra and Alejandra Gonzalez-Beltran and Claudiu Farcas and Todd Johnson and Ron Margolis and George Alter and Sansone, {Susanna Assunta} and Fore, {Ian M.} and Lucila Ohno-Machado and Grethe, {Jeffrey S.} and Hua Xu",

note = "Funding Information: This project is funded by grant U24AI117966 from the NIH National Institute of Allergy and Infectious Diseases as part of the Big Data to Knowledge program. We thank all members of the bioCADDIE community for their valuable input on the overall project. Funding Information: The biomedical and healthCAre Data Discovery Index Ecosystem (bioCADDIE) project,2 funded by the National Institutes of Health (NIH) via the Big Data to Knowledge program, is focused on the discovery of biomedical datasets. Since its start, researchers, service providers, and knowledge experts around the globe have participated in various aspects of bioCADDIE, such as working groups, pilot projects, and dataset retrieval challenges (https://biocaddie.org/). To instantiate the concepts and recommendations developed by this large community, bioCADDIE developed a prototype data discovery index (DDI) named DataMed, which collects and indexes metadata from broad types of biomedical datasets of interest from heterogeneous sources and makes them searchable through a web-based interface.3 We believe that metadata from diverse datasets can be mapped to a unified representation model, thus enabling more efficient search across domain-specific repositories and making data more discoverable by users. Further details and discussion of the motivations for building DataMed are available here.3 DataMed is available as an open source package, to allow the research community to leverage its technologies to build additional biomedical data indexes. This article describes technical details about developing DataMed, including its metadata ingestion/indexing pipeline and search engine functionalities. Publisher Copyright: {\textcopyright} The Author 2018. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved.",

year = "2018",

month = mar,

day = "1",

doi = "10.1093/jamia/ocx121",

language = "English (US)",

volume = "25",

pages = "300--308",

journal = "Journal of the American Medical Informatics Association",

issn = "1067-5027",

publisher = "Oxford University Press",

number = "3",

}

TY - JOUR

T1 - DataMed - an open source discovery index for finding biomedical datasets

AU - Chen, Xiaoling

AU - Gururaj, Anupama E.

AU - Ozyurt, Burak

AU - Liu, Ruiling

AU - Soysal, Ergin

AU - Cohen, Trevor

AU - Tiryaki, Firat

AU - Li, Yueling

AU - Zong, Nansu

AU - Jiang, Min

AU - Rogith, Deevakar

AU - Salimi, Mandana

AU - Kim, Hyeon eui

AU - Rocca-Serra, Philippe

AU - Gonzalez-Beltran, Alejandra

AU - Farcas, Claudiu

AU - Johnson, Todd

AU - Margolis, Ron

AU - Alter, George

AU - Sansone, Susanna Assunta

AU - Fore, Ian M.

AU - Ohno-Machado, Lucila

AU - Grethe, Jeffrey S.

AU - Xu, Hua

N1 - Funding Information: This project is funded by grant U24AI117966 from the NIH National Institute of Allergy and Infectious Diseases as part of the Big Data to Knowledge program. We thank all members of the bioCADDIE community for their valuable input on the overall project. Funding Information: The biomedical and healthCAre Data Discovery Index Ecosystem (bioCADDIE) project,2 funded by the National Institutes of Health (NIH) via the Big Data to Knowledge program, is focused on the discovery of biomedical datasets. Since its start, researchers, service providers, and knowledge experts around the globe have participated in various aspects of bioCADDIE, such as working groups, pilot projects, and dataset retrieval challenges (https://biocaddie.org/). To instantiate the concepts and recommendations developed by this large community, bioCADDIE developed a prototype data discovery index (DDI) named DataMed, which collects and indexes metadata from broad types of biomedical datasets of interest from heterogeneous sources and makes them searchable through a web-based interface.3 We believe that metadata from diverse datasets can be mapped to a unified representation model, thus enabling more efficient search across domain-specific repositories and making data more discoverable by users. Further details and discussion of the motivations for building DataMed are available here.3 DataMed is available as an open source package, to allow the research community to leverage its technologies to build additional biomedical data indexes. This article describes technical details about developing DataMed, including its metadata ingestion/indexing pipeline and search engine functionalities. Publisher Copyright: © The Author 2018. Published by Oxford University Press on behalf of the American Medical Informatics Association. All rights reserved.

PY - 2018/3/1

Y1 - 2018/3/1

N2 - Objective: Finding relevant datasets is important for promoting data reuse in the biomedical domain, but it is challenging given the volume and complexity of biomedical data. Here we describe the development of an open source biomedical data discovery system called DataMed, with the goal of promoting the building of additional data indexes in the biomedical domain. Materials and Methods: DataMed, which can efficiently index and search diverse types of biomedical datasets across repositories, is developed through the National Institutes of Health-funded biomedical and healthCAre Data Discovery Index Ecosystem (bioCADDIE) consortium. It consists of 2 main components: (1) a data ingestion pipeline that collects and transforms original metadata information to a unified metadata model, called DatA Tag Suite (DATS), and (2) a search engine that finds relevant datasets based on user-entered queries. In addition to describing its architecture and techniques, we evaluated individual components within DataMed, including the accuracy of the ingestion pipeline, the prevalence of the DATS model across repositories, and the overall performance of the dataset retrieval engine. Results and Conclusion: Our manual review shows that the ingestion pipeline could achieve an accuracy of 90% and core elements of DATS had varied frequency across repositories. On a manually curated benchmark dataset, the DataMed search engine achieved an inferred average precision of 0.2033 and a precision at 10 (P@10, the number of relevant results in the top 10 search results) of 0.6022, by implementing advanced natural language processing and terminology services. Currently, we have made the DataMed system publically available as an open source package for the biomedical community.

AB - Objective: Finding relevant datasets is important for promoting data reuse in the biomedical domain, but it is challenging given the volume and complexity of biomedical data. Here we describe the development of an open source biomedical data discovery system called DataMed, with the goal of promoting the building of additional data indexes in the biomedical domain. Materials and Methods: DataMed, which can efficiently index and search diverse types of biomedical datasets across repositories, is developed through the National Institutes of Health-funded biomedical and healthCAre Data Discovery Index Ecosystem (bioCADDIE) consortium. It consists of 2 main components: (1) a data ingestion pipeline that collects and transforms original metadata information to a unified metadata model, called DatA Tag Suite (DATS), and (2) a search engine that finds relevant datasets based on user-entered queries. In addition to describing its architecture and techniques, we evaluated individual components within DataMed, including the accuracy of the ingestion pipeline, the prevalence of the DATS model across repositories, and the overall performance of the dataset retrieval engine. Results and Conclusion: Our manual review shows that the ingestion pipeline could achieve an accuracy of 90% and core elements of DATS had varied frequency across repositories. On a manually curated benchmark dataset, the DataMed search engine achieved an inferred average precision of 0.2033 and a precision at 10 (P@10, the number of relevant results in the top 10 search results) of 0.6022, by implementing advanced natural language processing and terminology services. Currently, we have made the DataMed system publically available as an open source package for the biomedical community.

KW - Data discovery index

KW - Dataset

KW - Information dissemination

KW - Information storage and retrieval

KW - Metadata

UR - http://www.scopus.com/inward/record.url?scp=85043344475&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85043344475&partnerID=8YFLogxK

U2 - 10.1093/jamia/ocx121

DO - 10.1093/jamia/ocx121

M3 - Article

C2 - 29346583

AN - SCOPUS:85043344475

SN - 1067-5027

VL - 25

SP - 300

EP - 308

JO - Journal of the American Medical Informatics Association

JF - Journal of the American Medical Informatics Association

IS - 3

ER -

DataMed - an open source discovery index for finding biomedical datasets

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this