The aim of the crosscut metadata model is to provide a uniform encoding of metadata obtained from the following DCPPC data sources:
- AGR - Alliance of Genome Resources
- GTEx - Genotype-Tissue Expression project
- TOPMed - Trans-Omics for Precision Medicine
The metadata model specifies how the various metadata will be transformed into a uniform representation, whereas the metadata model instance is the transformed representation itself. The metadata model is based on a JSON-LD encoding of DATS, the DatA Tag Suite data model developed through the Big Data To Knowledge (BD2K) initiative to support dataset discoverability. Several extensions to DATS, along with schema.org and OBO Foundry-based JSON-LD context files, have been developed to support the crosscut metadata model. These extensions and context files can be found in the DATS GitHub repository at https://github.com/datatagsuite. For more information on how the 3 main datasets' metadata are encoded in DATS, see the relevant sections below.
For those familiar with RDF and SPARQL, one of the quickest ways to gain familiarity with the metadata model and the current instance is to examine and run the example SPARQL queries. To do this, first read and follow the instructions in the next section, entitled "Downloading the Public Crosscut Metadata Model Instance". Then retrieve and run the scripts in the following directory, as directed in the README.md therein:
https://github.com/dcppc/crosscut-metadata/tree/master/sparql/v0.6
The crosscut metadata model instance, which is essentially a small set of JSON-LD files, is distributed as a BDBag. BDBags for all current public releases can be found in the releases/ subdirectory. Each BDBag is a gzipped tar file that can be retrieved, extracted and uncompressed with standard Unix or Mac OS command line utilities. On a Mac, for example, the latest (as of this writing) v0.6 release can be retrieved and uncompressed with the following commands:
$ curl -s -O 'https://raw.githubusercontent.com/dcppc/crosscut-metadata/master/releases/KC7-crosscut-metadata-v0.6.tgz'
$ tar xzvf KC7-crosscut-metadata-v0.6.tgz
x KC7-crosscut-metadata-v0.6/
x KC7-crosscut-metadata-v0.6/tagmanifest-md5.txt
x KC7-crosscut-metadata-v0.6/bagit.txt
x KC7-crosscut-metadata-v0.6/bag-info.txt
x KC7-crosscut-metadata-v0.6/tagmanifest-sha256.txt
x KC7-crosscut-metadata-v0.6/manifest-md5.txt
x KC7-crosscut-metadata-v0.6/data/
x KC7-crosscut-metadata-v0.6/data/datasets/
x KC7-crosscut-metadata-v0.6/data/datasets/TOPMed_phs000951_phs000946_phs001024_wgs_public.jsonld
x KC7-crosscut-metadata-v0.6/data/datasets/GTEx_v7_public.jsonld
x KC7-crosscut-metadata-v0.6/data/datasets/AGR_MGI_RGD.jsonld
x KC7-crosscut-metadata-v0.6/data/docs/
x KC7-crosscut-metadata-v0.6/data/docs/RELEASE_NOTES
x KC7-crosscut-metadata-v0.6/data/docs/ChangeLog
x KC7-crosscut-metadata-v0.6/manifest-sha256.txt
After uncompressing the DATS JSON-LD files can be found in KC7-crosscut-metadata-v0.6/data/datasets:
$ ls -al KC7-crosscut-metadata-v0.6/data/datasets/
total 1457784
drwxr-xr-x 5 jcrabtree staff 160 Sep 14 19:47 .
drwxr-xr-x 4 jcrabtree staff 128 Sep 14 19:56 ..
-rw-r--r-- 1 jcrabtree staff 421546018 Sep 14 19:47 AGR_MGI_RGD.jsonld
-rw-r--r-- 1 jcrabtree staff 323335853 Sep 14 19:47 GTEx_v7_public.jsonld
-rw-r--r-- 1 jcrabtree staff 1497331 Sep 14 19:47 TOPMed_phs000951_phs000946_phs001024_wgs_public.jsonld
Note that if the bdbag utility is run to extract the BDBag after unpacking it then the data portion
of the above path will not be present.
The script to build the public crosscut metadata model instance is called make-crosscut-instance-bdbag.sh
and can be found in the top level of this repository:
https://github.com/dcppc/crosscut-metadata/blob/master/make-crosscut-instance-bdbag.sh
The script contains the commands to perform the DATS metadata conversion for each of the currently supported data (sub)sets, but as the comments in the file indicate, the metadata flat files for each of the data sources must first be downloaded to the current directory:
For AGR the following resources are needed to generate the current instance, which includes reference annotation for mouse and rat:
AGR filtered ortholog .tsv file:
- alliance-orthology-july-19-2018-stable-1.6.0-v4.tsv
AGR Basic Gene Information (BGI), GFF3, and disease_json files:
- MGI_1.0.4_BGI.json
- MGI_1.0.4_disease.json
- MGI_1.0.4_GFF.gff
- RGD_1.0.4_BGI.json
- RGD_1.0.4_disease.json
- RGD_1.0.4_GFF.gff
For GTEx the following two files are needed from https://www.gtexportal.org/home/datasets:
GTEx_v7_Annotations_SubjectPhenotypesDS.txt
GTEx_v7_Annotations_SampleAttributesDS.txt
In addition the public dbGaP variable summaries and data dictionaries should be downloaded
from the following URL into a local directory named phs000424.v7.p2:
ftp://ftp.ncbi.nlm.nih.gov/dbgap/studies/phs000424/phs000424.v7.p2/pheno_variable_summaries/
Finally, the manifest and id dump files from the DCPPC data-stewards GitHub repository are required:
git clone https://github.com/dcppc/data-stewards.git
For the example TOPMed studies (phs001024, phs000951, and phs000179) the public TOPMed dbGaP
variable summaries and data dictionaries should be downloaded from the following URLs into
local directories named phs001024.v3.p1, phs000951.v2.p2, and phs000179.v5.p2,
respectively:
ftp://ftp.ncbi.nlm.nih.gov/dbgap/studies/phs001024/phs001024.v3.p1/pheno_variable_summaries/
ftp://ftp.ncbi.nlm.nih.gov/dbgap/studies/phs000951/phs000951.v2.p2/pheno_variable_summaries/
ftp://ftp.ncbi.nlm.nih.gov/dbgap/studies/phs000179/phs000179.v5.p2/pheno_variable_summaries/
In order to run the part of the script that creates the BDBag, the bdbag command-line utility
must be installed like so:
pip install bdbag
The procedure for downloading the controlled access version of the metadata model can be found at /docs/controlled_access_download.md
The script mentioned above, make-crosscut-instance-bdbag.sh, also contains example commands
(which are commented out by default) showing how to generate DATS JSON for the controlled access
metadata available from dbGaP for the GTEx and TOPMed studies. Download the controlled access
dbGaP files to a local directory with the appropriate access controls, and then tell the
conversion script where to find the public and controlled access metadata files, as shown in
the script.
Modified versions of the public DATS JSON-LD files have all been validated against the latest (as of this writing) version of DATS from
https://github.com/datatagsuite
using the validator in
https://github.com/datatagsuite/dats-tools.
This repository contains an extension of DATS that is specific to this project and the crosscut metadata model instance JSON-LD files will NOT necessarily validate against the current DATS release found at https://github.com/biocaddie/WG3-MetadataSpecifications. Any changes to the DATS JSON files are checked against the validator before creating a new release of the metadata model instance.
Validation was performed using modified versions of the JSON-LD files because the DATS validator
does not yet support circular links or arbitrary id references in the JSON-LD files. Circular links
are removed and id references are dereferenced for validation and then added back in to create the
final released JSON-LD files. Comments in the make-crosscut-instance-bdbag.sh script describe
how this is done.
This section describes how the three datasets are currently encoded in DATS and discusses some of the tradeoffs and shortcomings of the encoding. The encoding is by no means set in stone and the process of refining and improving it is still ongoing. Concomitant adjustments are also being made to the DATS model in some cases to facilitate the encoding of some aspects of the metadata.
Similar to the GTEx and TOPMed encodings, the top level of the AGR JSON-LD structure is a DATS
Dataset that represents all Alliance data. Below that is a set of 2nd-level Dataset entities,
linked by the hasPart property. Those 2nd-level Datasets correspond to the individual reference
genomes from the model organism databases that comprise the Alliance. The similarity with the
GTEx and TOPMed encodings ends at this point: below each 2nd-level Dataset is an array of
MolecularEntity objects, linked to it by the isAbout property. These MolecularEntities
represent the genes and/or pseudogenes in each reference genome and make use of the new (as of v0.6)
DATS GenomeLocation schema to specify the genomic sequence location(s) associated with each.
The DATS alternateIndentifiers of each gene link it to the corresponding NCBI_HomoloGene group
and/or human gene(s). The relatedEntities property provides explicit links to orthologs (i.e.,
represented as a DATS MolecularEntity) as well as disease and/or phenotype relationships.
The following ER diagram illustrates the overall structure:
At the top level of both the GTEx and TOPMed encoding is a DATS Dataset that represents the project
(GTEx) or program (TOPMed) as a whole. This top level Dataset is linked by the hasPart property
to an array of DATS Datasets, each of which represents a specific study (in the general sense, since
Study is a DATS entity in its own right.) For both TOPMed and GTEx each of the 2nd-level Dataset
entities corresponds to a dbGaP study. For GTEx there is only one such study, phs000424, but for
TOPMed there are several, some of which are "parent" studies (i.e., they existed in dbGaP before
becoming part of the TOPMed program) and some of which are "TOPMed" studies (i.e., the ones with
which the TOPMed WGS data are associated.) Below these 2nd level Datasets that represent studies
are 3rd level Datasets that represent the individual data files/products produced by the study.
Below the GTEx dbGAP study, for example, the DATS hasPart property links to an array of DATS
Datasets, each of which represents one of the controlled access RNA-Seq or WGS data files.
Each of the second level DATS Dataset objects is in turn linked to an array of DATS Material objects
by the isAbout property. Each of those Materials represents an RNA extract used in the RNA-Seq
protocol or, in the case of WGS sequence data, a DNA extract used in WGS sequencing.
In DATS a Material may be linked to one or more additional Material objects via the derivesFrom
property. In the GTEx and TOPMed encoding each RNA (or DNA) extract Material is linked first (via
derivesFrom) to a Material that represents a biological sample from a particular body site. That
biological sample Material is further linked (also via derivesFrom) to a Material that represents
the individual human donor/subject, as shown in the following ER diagram:
In the public version of the GTEx DATS encoding all of the human subjects, samples, and RNA and DNA extracts are represented, but some of the phenotype and/or sample data may be limited. For example, instead of specifying each subject's exact age, only an "Age range" (e.g,. "60-69") is provided.
In all of the dbGaP-based public TOPMed DATS JSON encoding the entities that represent human subjects, samples, and RNA/DNA extracts are dummy entities produced by picking the most commonly-occurring values from the dbGaP variable summary reports.
The following simplified ER diagram illustrates the overall structure of the TOPMed encoding:
And this (very similar) one does the same for the GTEx encoding:
