Presentation on theme: "THE PUBLISHER'S PERSPECTIVE AND THE F1000 APPROACH TO VERSIONING Rebecca Lawrence, PhD Publisher, F1000 Research"— Presentation transcript:
THE PUBLISHER'S PERSPECTIVE AND THE F1000 APPROACH TO VERSIONING Rebecca Lawrence, PhD Publisher, F1000 Research
FACULTY OF 1000 Faculty of 1000 SubscriptionOpen Access
F1000PRIME, F1000TRIALS AND F1000POSTERS F1000Prime Launched in 2002, from founders of Current Opinion journals and BMC Across biology and medicine 10,000 experts that identify and evaluate the most important articles 1,500 new recommendations per month; >130,000 total so far F1000Trials (Beta) Comprehensive listing and commentary from ~500 journals Covers RCTs, early phase trials, systematic reviews and meta-analyses F1000Posters Open Access repository of posters and slides in biology and medicine > 4000 so far; > 200 are evaluated F1000Research (Beta) F1000.com
THE ISSUES AND PROBLEMS WITH VERSIONING There may be several ‘versions’ of a given article available: Publisher PMC Indexers Author version on their website Institutional repository version etc One of the roles of the publisher is to maintain a clear formal ‘version of record’ for an article. This means nothing should change following publication
HOWEVER... This is not how science is really conducted Some publishers are starting to explore versioned articles Authors can amend their articles post-publication What happens when the data associated with the article changes over time It might be say an annotated record (e.g. SwissProt) that changes as new publications come out It might be ongoing time-series data It might be new data being added as the ‘n’ number gets higher What happens when an error is spotted that needs correcting (CrossMark)? What constitutes enough of a change to be a new version, or indeed a new article?
F1000RESEARCH An Open Access journal that redefines scientific publishing by offering: 1.Immediate publication (within one week) 2.Completely transparent peer review, post publication 3.Full data deposition and sharing 4.Acceptance of all scientifically sound articles, including single findings, case reports, protocols, null/negative results etc.
F1000RESEARCH: OTHER KEY FEATURES Covers all biology and medicine ‘Gold’ Open Access Creative Commons CC-BY licences as default; CC0 licence for data as default Large (230+), very senior Advisory Panel (e.g. Sir Tim Hunt, Pippa Marrack, Steven Hyman, Alan Schechter, Janet Thornton, Phil Bourne) Large (1000+) Editorial Board of very senior names across life sciences Beta-launched in Jul 2012; formal launch on our own platform in Jan 2013
THE PUBLISHING PROCESS Publication Pre- publica tion check Data repository Open referee process User commenting Author revises article Article submitted CrossRef DOI DataCite DOIs CrossRef DOI DataCite DOIs Data repository
A TWO-STEP PEER REVIEW PROCESS FIRST: Rapid ‘seems ok’ stamp – 3 options: Approved (like approved or minor revisions) Approved with Reservations (like major revisions) Not Approved (like reject) SECOND: Referee comments All open and signed; Focus is on whether the work is scientifically sound, not on novelty/interest etc
CITATION Citation includes details of: Version number Referee status Shortened URL to latest referee status Indexed once it receives 2 positive reviews (‘Approved’, or ‘Approved with Reservations’ statuses). If it receives ONLY negative reviews (2+), then not indexed and excluded from the default search on the site. DOIs are versioned e.g /f1000research v2 (so similar to data citation versioning in EBI etc)
VERSIONED URLS Version 2: Version 1: Core URL: Latest article version
INDEXING – SCOPUS AND EMBASE Will combine the citations from all versions To do this, they will remove the ‘vx’ at the end of the DOI and the text in square brackets at the end of the article title As most fields can change (author list, affiliations abstract, references, etc), core identifying fields will be: First author Article number Title To avoid multiple versions on searching/browsing, they will only keep the latest version They will then send users to the F1000Research site where they will be informed of previous versions should they wish to see them The British Library are planning to use the same approach; we are still discussing with PMC/PM but it will likely be similar again.
INTERESTING ISSUES THAT COME OUT OF THIS 1.If you add an author to a new version, is it right that they benefit from citations of the original paper? We decided yes, that is the only way to do this sensibly, certainly as a starting point 2.If Version 1 was published in December 2012, and Version 2 is published in February 2013, should the citation for Version 2 remain a 2012 publication? We decided yes otherwise you lose priority if your article now looks to be published after a competitors’ paper that came out in say January Should the title be able to change? We decided that it should not, once the article reaches the indexed stage: o We will check with the authors prior to sending to indexers if they wish to change their title following referee comments o Beyond that, a change in title suggests a new article is required 4.The article (html and PDF) are added to every time a new referee report or comment comes in, but the DOI does not change – should it?
F1000R: EMBED WIDGETS Through collaboration with FigShare, and soon Dryad as well, we provide an ‘embed widget’ that: Enables viewing of the data without needing to leave the article Provides viewers for data files Can preview large datasets before deciding whether to download Provides information on views, shares and downloads Datasets get legends and DOIs so they can be independently cited Data on the repository site are then linked back to our DOI Through CrossMark, the DOIs are linked up
F1000R: INTERESTING TYPES OF ARTICLE RECEIVED Data-only articles Linked data and research articles Ongoing updated articles Time-series data Data added in later versions e.g. after referee request or higher n number Re-analyses of the current state of affairs on an annual basis e.g. The MTBBASE-GOA dataset, which contains manually curated Gene Ontology annotations of the M. tuberculosis genome
SUMMARY The enabling of versioning has many uses and allows articles to more closely align with how research is actually conducted Versioning is also very valuable as publishers increasingly integrate articles with data The indexers are starting to address the many complications caused by versioning Many data repositories already have mechanisms for versioning that we can learn from. Thank