1.
Better Metadata for
Better Data Reuse
in Biomedicine:
The CEDAR Project
1
1

2.
A Metadata Ecosystem
• HIPC investigators perform experiments in
human immunology.
• HIPC Standards Working Group makes metadata
templates to encourage uniform experiment
data annotation.
• HIPC investigators fill out HIPC metadata using
the metadata templates.
• ImmPort stores HIPC data (and metadata) in its
public repository for data discovery and use.
• Researchers use resulting metadata resources. 2
2

3.
Example entity–relationship diagram for describing
metadata for annotating multiplex bead array assays
(e.g., Luminex)
HIPC centers
submit data
directly to the
NIAID ImmPort
data repository
using detailed,
experiment-
specific templates
3
3

4.
4
4

5.
MIAME and other minimal
information checklists
address much of the
biomedical community
• Processes to apply them
are complicated
• Investigators see even
“minimal” information as
burdensome
5
5

6.
6
6

7.
The CEDAR Vision
• End-to-end improvement in
biomedical metadata
• Compatibility with best known
practices and standards
• Use of existing semantic and
data structure resources
• A simple ‘life of metadata’ for
users
7
7

8.
How can we make
metadata authoring better?
• Let any user build “smart” metadata templates
• Use standard value sets to fill in the blanks
• Make great user interfaces and powerful APIs
• Use machine learning to predict and validate
metadata entries
• Provide easy upload of metadata (and data)
• Use text processing to acquire metadata from
written descriptions of the experiment
8
8

9.
CEDAR’s approach to metadata
9
Training Activities
9

10.
CEDAR technology
• Interfaces for users
– Create metadata template fields based on controlled terms
– Assemble them into composite templates (a la forms)
– Fill out templates with experimental metadata
• A repository of metadata
– Enables metadata discovery
– Guides predictive entry of new metadata
– Informs creation of templates
• Links to unique, rigorous encodings and open data
– Metadata fields are well described (a la CDEs)
– Metadata are encoded using appropriate terms
– References can be found and clearly cited
10
10

11.
11
Template Specifiers Research Scientists
11

12.
12
12

13.
Template Designer: create templates
13

14.
Metadata Editor: collect metadata
14

15.
15

16.
16

17.
17

18.
18

19.
NCBI-AIRR Submission Workflow
19

20.
LINCS Data Standards in CEDAR Workflow
20

21.
Making Authoring Metadata Easier
• A library of standard templates like MIAME, in a
sharable metadata exchange format
• Direct access to biomedical ontologies through
BioPortal, and simple selection of concepts
• Guidance on the frequency of menu selections
• Predictive data entry whenever possible
• Direct upload to target repositories
21

22.
Using A Principled Approach
• All knowledge artifacts managed as first-class entities
• On-the-fly user interfaces and drop-down menus
• All software components have well-defined APIs
• All content represented in JSON-LD
• User-friendly interfaces minimize entry time
22

23.
Resources
Web site: http://metadatacenter.org
Workbench: https://cedar.metadatacenter.org
Technical site: https://metadatacenter.github.io
API: https://resource.metadatacenter.org/api/
23

24.
JSON Schema + JSON-LD JSON-LD
Using JSON Schema and JSON-LD for
CEDAR Template Model
24

25.
What is JSON Schema?
• Technology for describing and validating the
structure of JSON documents
• Provides a structural description of any
JSON document
• JSON documents that are specified with
JSON Schema can be structurally validated
against their associated schemas
• Analogous to XML Schema
25

26.
What is JSON-LD?
• A lightweight syntax to serialize Linked Data in JSON
• Allows existing JSON to be interpreted as Linked Data
with minimal changes
• JSON-LD is primarily intended to allow:
– using Linked Data in Web-based programming
environments
– building interoperable Web services
– storing Linked Data in JSON-based storage engines
• Core contribution: add semantics to JSON documents
• W3C Recommendation: https://www.w3.org/TR/json-ld/
26

27.
{
"$schema": "http://json-schema.org/draft-04/schema#",
"@type": "https://repo.metadatacenter.org/core/Template",
"@id": "https://repo.metadatacenter.org/templates/434334",
"title": ”Study",
"description": ”Study template",
"type": "object",
"_ui": {...},
"properties": {
"title": {...},
”description": {...},
”principalInvestigator": {...}
},
"required": ["title", "description",
"principalInvestigator"]
"additionalProperties": false
}
Using JSON Schema to Define Template
Structure
27

28.
{
"title": { "@value": "Immune biomarkers study" },
"description": { "@value": "Immune biomarkers …" },
"principalInvestigator": {
"name": { "@value": "Dr. P.I" },
"institution": {
"name": { "@value": "Stanford" },
"zip": { "@value": "94305" }
}
}
}
Using JSON-LD to add Semantics to
Metadata Instances
28

29.
{
"@type": "http://semantic-dicom.org/dcm#Study",
"@id": "https://repo.metadatacenter.org/template_instances/55417",
"@context": {
"title": "https://schema.org/title",
"name": "https://schema.org/name",
"description": "https://schema.org/description",
"zip": "https://schema.org/postalCode",
"pi": "https://myschema.org/property/hasPI",
"institution": "https://myschema.org/property/hasInstitution"
},
"title": { "@value": "Immune biomarkers study" },
"description": { "@value": "Immune biomarkers …" },
”principalInvestigator": {
"@type": "https://schema.org/Person",
"@id": "https://repo.metadatacenter.org/template_elements/557",
"name": { "@value": "Dr. P.I" },
"institution": {
"@type": "https://schema.org/Organization",
"@id": "https://repo.metadatacenter.org/template_elements/37",
"name": { "@value": "Stanford" },
"zip": { "@value": "94305" }
} } }
Using JSON-LD to add Semantics to
Metadata Instances - II
29

30.
Metadata Instances as RDF Graph
30

31.
Generic CEDAR Workflow
Metadata
External
31

32.
Workflow for LINCS
32

33.
Workflow for LINCS
33

34.
Workflow for LINCS
34

35.
Workflow for AIRR in NCBI
35

36.
Workflow for AIRR in NCBI
36

37.
Workflow for AIRR in NCBI
37

38.
Additional Applications
38

39.
Projects Under Discussion
NCI Proteomics Data Commons
39