| title | title_short | tags | authors | affiliations | date | bibliography | git_url | event | group | authors_short | header-includes | ||||||||||||||||||||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Metadata for BioHackrXiv Markdown publications |
Metadata for BioHackrXiv |
|
|
|
15 November 2020 |
paper.bib |
Elixir2020 |
BioHackrXiv group |
Pjotr Prins, Tazro Otha, Leyla Garcia Castro & Toshiaki Katayama |
\usepackage{caption}
\captionsetup[figure]{
name=,labelsep=none,labelformat=empty}
|
https://biohackrxiv.org/ is a scholarly publication service for BioHackathons and Codefests where papers are generated from Markdown templates where the header is a YAML/JSON record that includes the title, authors, affiliations and tags. The idea originated from the Markdown layout used in the Journal of Open Source Software (JOSS)[@JOSS]. Templates are provided as an example. A minimal metadata header used for this publication has the following layout
\scriptsize
---
title: 'Metadata for BioHackrXiv Markdown publications'
title_short: 'Metadata for BioHackrXiv'
tags:
- metadata RDF pre-prints
authors:
- name: Pjotr Prins
orcid: 0000-0002-8021-9162
affiliation: 1
- name: Tazro Ohta
orcid: 0000-0003-3777-5945
affiliation: 2
- name: Leyla Garcia Castro
affiliation: 3
orcid: 0000-0003-3986-0510
- name: Toshiaki Katayama
orcid: 0000-0003-2391-0384
affiliation: 2
affiliations:
- name: Department of Genetics, Genomics and Informatics, The University of Tennessee Health Science Center, Memphis, TN, USA.
index: 1
- name: Database Center for Life Science, Joint Support-Center for Data Science Research, Research Organization of Information and Systems, Japan
index: 2
- name: Knowledge Management Group, at ZBMED Information Centre for Life Sciences
index: 3
date: 15 November 2020
bibliography: paper.bib
git_url: https://github.com/biohackrxiv/bhxiv-gen-pdf
event: Elixir2020
group: BioHackrXiv group
authors_short: Pjotr Prins, Tazro Otha, Leyla Garcia Castro & Toshiaki Katayama
---\normalsize
BioHackrXiv publications are hosted at the Open Science Framework (OSF), an open source website that connects and supports the research workflow, enabling scientists to increase the efficiency and effectiveness of their research. OSF is run by the non-profit Center for Open Science (COS) and charges for the handing out of digital object identifiers (DOIs). The OSF site provides support for basic editorial functions and displaying PDFs. OSF also provides a basic search infrastructure to find these 'preprint' publications.
Many projects in BioHackathons are about using FAIR data. The OSF site lacks a number of facilities, mostly related to metadata. For example, it is not easy or straightforward to list publications related to certain BioHackathons, e.g. `BioHackathon 2019 Japan'. Another problem is that there is no clear link between publications and the projects they refer to. It is necessary to read the PDF for that.
Because the current setup is lacking in the findable (F) and accessible (A) of FAIR, for the ELIXIR BioHackathon 2020, we decided to add an additional service that provides a SPARQL endpoint for queries and some simple HTML output that can be embedded in a BioHackathon website.
To connect the main publication with its source and related Markdown we maintain a list in a Git repository that needs to be updated with every publication. This file contains hard links we can not easily acquire in other ways. E.g.
\scriptsize
papers:
- id: https://biohackrxiv.org/km9ux/
doi: https://doi.org/10.37044/osf.io/km9ux
markdown: https://raw.githubusercontent.com/journal-of-research-objects/
Example-BioHackrXiv-Paper/master/paper.md\normalsize
The Markdown link should be able to fetch the parsable Markdown file and fetch the contained header metadata. Any conflicting metadata added by papers.yaml file overrides the metadata in the Markdown file.
Some metadata, such as the DOI, is presented as HTML on the OSF web pages. In future work this data can be scraped to add to the metadata stored in papers.yaml.
After compiling the metadata we transform that into Resource Description Framework (RDF), the language of the Semantic Web, using a small subset of the Dublin core and scholarly publication ontology from schema.org.
In the first version we generate RDF with Ruby ERB templates which means that validation happens in the source code. Some semantic enrichment includes URIs for the BioHackathons themselves. An example for one paper:
\scriptsize
<http://2019.biohackathon.org/> schema:description "NBDC/DBCLS BioHackathon, Fukuoka, Japan, 2019"@en ;
schema:name "Japan2019" ;
a schema:Event .
<https://biohackrxiv.org/km9ux/> dc:title "Logic Programming Working Group"@en ;
schema:sameAs <https://doi.org/10.37044/osf.io/km9ux> ;
schema:url <https://raw.githubusercontent.com/journal-of-research-objects/
Example-BioHackrXiv-Paper/master/paper.md> ;
bhx:Event <http://2019.biohackathon.org/> ;
a schema:CreativeWork .
<https://biohackrxiv.org/km9ux/> dc:contributor "Chris Mungall"@en .
<https://biohackrxiv.org/km9ux/> dc:contributor "Hirokazu Chiba"@en .
<https://biohackrxiv.org/km9ux/> dc:contributor "Shuichi Kawashima"@en .
<https://biohackrxiv.org/km9ux/> dc:contributor "Yasunori Yamamoto"@en .
<https://biohackrxiv.org/km9ux/> dc:contributor "Pjotr Prins"@en .
<https://biohackrxiv.org/km9ux/> dc:contributor "Nada Amin"@en .
<https://biohackrxiv.org/km9ux/> dc:contributor "Deepak Unni"@en .
<https://biohackrxiv.org/km9ux/> dc:contributor "<nobr>William E. Byrd</nobr>"@en .
\normalsize
Where the embedded HTML should be filtered out. The full current RDF can be viewed here.
In future work we may introduce ShEx to validate entries[@shex]. Also the scholarly annotation can be expanded. For example, we could parse the number of downloads and other information from the OSF website and transform that into RDF. We can also add links to indexing services, such as PubMed Central and Google Scholar, when they have added BioHackrXiv publications/reports.
Once the RDF is uploaded into a triple store, such as Virtuoso, it is possible to write SPARQL queries that return records in a JSON format that can be parsed inside a (BioHackaton) website. For example, the following queries lists all BioHackathons at time of writing:
\scriptsize
prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
prefix dc: <http://purl.org/dc/terms/>
prefix bhx: <http://biohackerxiv.org/resource/>
prefix schema: <https://schema.org/>
SELECT ?url ?name ?descr
FROM <https://biohackrxiv.org/graph>
WHERE {
?url schema:name ?name .
?url schema:description ?descr
}
BioHackathon Europe, Paris, France, 2019
BioHackathon Europe 2020 Online
NBDC/DBCLS BioHackathon, Fukuoka, Japan, 2019
NBDC/DBCLS BioHackathon, Matsue, Japan, 2018
Virtual BioHackathon Covid-2020\normalsize
To list all papers for one BioHackathon (Virtual BioHackathon Covid-2020):
\scriptsize
SELECT ?title ?url
FROM <https://biohackrxiv.org/graph>
WHERE {
?bh schema:name "Covid2020" .
?url bhx:Event ?bh ;
dc:title ?title .
}
Global analysis of human SARS-CoV-2 infection and host-virus interaction
https://biohackrxiv.org/b4zkp/
Comparison of SARS-CoV-2 variants with INSaFLU and galaxyproject
https://biohackrxiv.org/9d3cz/
Characterization of Potential Drug Treatments for COVID-19 using Twitter
https://biohackrxiv.org/cu2s9/
Determining a novel feature-space for SARS-COV2 Sequence data
https://biohackrxiv.org/xt7gw/\normalsize
After adding contributors to RDF we can count all contributors
\scriptsize
prefix bhx: <http://biohackerxiv.org/resource>
SELECT DISTINCT COUNT(?contributor)
FROM <https://BioHackrXiv.org/graph>
WHERE {
?bh <http://purl.org/dc/terms/contributor> ?contributor
}
73 contributors\normalsize
See the README file in the GitHub repo for a recent SPARQL endpoint that supports these queries.
Once the data is in a triple store database and can be queried with SPARQL we can embed the results in a web page. We created a small Ruby Sinatra web service in https://github.com/biohackrxiv/bhxiv-metadata/web. This mini service fetches papers by BioHackathon, e.g. http://server/list?bh=Covid2020 will list all papers related to the Covid2020 BioHackathon. The keywords are listed in etc/events.yaml in the source tree. In our example a SPARQL query is executed in the backend and the JSON is transformed into embeddable HTML. As an alternative, rather than using server side parsing, a page could also be generated by JavaScript using SPARQL directly using, for example, the SPARQL.js module.
As we are already running a webserver for http://preview.biohackrxiv.org/ it made sense to embed some functionality in the Ruby Sinatra driven webserver. This code can be found here.
Websites and Git repositories do not live forever. Therefore we decided to cache the papers and metadata locally in the Git repository. For now we can fetch files and simply store them. When the fetch stops working we fall back on the cached version. One advantage of this approach is that it is easy to track updates to upstream repositories.
As part of the ELIXIR BioHackathon 2020 we created a metadata resource for BioHackrXiv, a prepublishing service hosted on osf.io that allows for citeable BioHackathon reports.
We added metadata in RDF with information on the BioHackathons, papers, repositories, contributors and tags. This metadata can be expanded and that can easily be done by modifying the source code in below repository.
Even though osf.io does not provide all the functionality we require for BioHackrXiv, we are able to work around limitations and our functionality may be merged or linked into the main https://biohackrxiv.org/ website in the future. For now, we will host a separate web server we operate ourselves and possibly add IPFS permanent URLs that can be used in addition to DOI to reference a document with a unique identifier [@IPFS].
We invite contributions to parsing and adding relevant metadata to BioHackrXiv RDF. Simply do a pull request on https://github.com/biohackrxiv/bhxiv-metadata. The web server examples are hosted on https://github.com/biohackrxiv/bhxiv-gen-pdf.
We thank the organizers of the ELIXIR BioHackathon 2020 for the event. We also thank Database Center for Life Science (DBCLS) for sponsoring the osf.io hosting of BioHackrXiv.