BLAH (Biomedical Linked Annotation Hackathon) series is organized to promote collaboration for sharing and linking resources for biomedical literature annotation and mining. Here, interested resources include corpora, annotation datasets, databases, software tools and services, terminologies, ontologies, graphics, movies, and so on. However, we hope they are all to be linked to each other, particularly through literature, as literature for a long time has played a central role for publication of scientific discoveries.
As a hackathon event, we are going to pay our efforts to implement things: each participant is encouraged to come with resources to share and to make links to other resources. Through collaboration during day time, those resources will become more available to the public in an interlinked way. During night time, a venue for creative discussions will be organized, which sometimes bring brilliant ideas. We solicit participation of researchers, engineers, and practitioners, who are interested in sharing and linking resources for biomedical literature mining.
BLAH6 is organized as a one-day symposium and a three-day hackathon. The symposium day consists of a series of keynote addresses and talks by participants. The hackathon focuses on projects proposed by the participants themselves.
In this year, we set "social media mining" as the special theme of BLAH6, and encourage potential participants to consider their contribution to that direction.
We invite proposals for symposium talks and associated projects. The idea is that the symposium talk presents a project idea, with the goal of recruiting other participants to work on it. Participants then choose which project(s!) to participate in, and the hackathon itself is built around those projects. (Other projects are welcome, but the probability of recruiting other participants to your project is higher if you present it on the symposium day.)
Consistent with BLAH's role as an open forum for work on linking resources for biomedical text mining, we solicit proposals for projects on resources, linking, biomedical text mining, and applications of those three in combination. Examples of good BLAH6 topics include:
- Annotated data sets
- Annotation tools
- Applications of linked annotations
- Usability testing
- User interface design issues
- Reproducibility issues
A submission should be 1-2 pages long (up to approximately 1,400 words or 1,000 words plus one figure) in PDF and should include:
- Author information
- Proposal type (see below)
- A description of the project
- The global motivation for the project
- Specific realistic goals for the period of the hackathon
- An explanation of how the hackathon environment would contribute to realization of the project
- A plan for evaluating the work
- Request for travel support, if it is needed
See below for specific criteria that you should address for proposals related to (1) annotated data sets, (2) annotation tools, and (3) applications.
Submissions have to be made through Easychair (https://easychair.org/conferences/?conf=blah6). Every submitted proposals will be reviewed by PC to determine their suitability for BLAH6. Selected proposals will be presented for the first day symposium. We expect successful proposals to find interested participants who will collaborate on them during the hackathon. After the hackathon, a journal special issue (see below) will be organized to publish outputs of BLAH6 which we expect to be extensions of the proposals (See below).
- Submission due: 29th (Friday) November, 2019
- Notification : 9th (Monday) December, 2019
Criteria for specific proposal types
Annotated data sets
If you have a data set of literature annotation (particularly to PubMed or PubMed Central articles), e.g. the NCBI disease corpus which contains disease annotations of 793 articles, you may want to make it interoperable with other annotations, to improve its own value and to find new applications of it. You can find other annotations contributed through previous BLAH at the PubAnnotation homepage.
Prospective contributors are asked to submit a description of the data sets (1-2 pages) which should include:
- name of the text annotation data
- the name and e-mail address of the contact person(s)
- how the base texts are collected
- how the annotations are developed, e.g., manually or automatically
- a small piece of example and a detailed description of it
- reference to relevant publications or web pages if any
- license policy (Creative Commons preferred)
Alternatively, and more ideally, a homepage may be prepared with descriptions of all the above aspects. In the case, simply the URL of the homepage can be submitted.
Expected types of annotations to be submitted may include, but not be limited to:
- Manual or automatic annotations
- Semantic or syntactic annotations
- Biological or clinical annotations
We refer here to tools for manual or automatic annotation. We solicit tools and systems for production of literature annotation which will willingly be publicly available. Prospective tools will be supported to be made inter-operable with other tools (mainly through REST API and JSON).
As an example, a way of linking an automatic annotation tool from PubAnnotation is described in this document.
Also, a way of making an annotation editor compatible with PubAnnotation is described in this document.
Expected type of contribution will include, but not limited to:
- visualization (e.g., TextAE)
- automatic annotation (e.g., TagTog)
- format conversion (e.g., simple_bioc)
Prospective contributors are asked to submit a description of the tools (1-2 pages) with reference to relevant publications if any.
Successful contributions must
- be web-based applications,
- guarantee free access without restriction to the public,
- open-source software is desired.
- provide RESTful APIs for programmable access, and
- provide comprehensive online documentation.
Alternatively, and more ideally, a homepage may be prepared with descriptions of all the above aspects. In that case, simply the URL of the homepage can be submitted.
Applications of linked literature annotation
To promote next generation text mining research based on the linked literature annotation resources, we also call for position papers to present innovative ideas of utilizing or extending the potential of the linked annotation resources.
Scope of interest may include, but not be limited to:
- applications which can be linked to biologically or medically meaningful discoveries
- integration of existing workflows
- tools for utilization of linked resources, e.g., analysis, visualization, conversion, mining
A submission may consist of up to 1000 words with figures and tables. A submission shall be a PDF file which includes all the figures and tables.
It is expected that successful proposals shall present
- proof-of-concept examples
- data sets to be used (if the data sets are developed by other groups, submitting a joint proposal is recommended)
For each accepted proposal, it is expected that at least one author shall attend the BLAH hackathon/symposium to present the work to seek potential collaboration with other meeting attendees (e.g. data providers).
Special issue of Genomics & Informatics
Soon after BLAH6, a special issue of the journal of Genomics & Informatics will be organized to publish outputs of BLAH6.
The journal is
- an open access journal,
- indexed by PubMed, and
- all the articles are immediately included in PubMed Central.
Participants to BLAH6 will be invited to submit manuscripts, based on their results from the hackathon.
Below is special offers for each article to be included in the special issue:
- the article processing charge will be waived,
- English proof reading service will be provided for free, and
- a sponsor travel support (approximately 250USD) will be awarded.