As a stimulus for thinking about today's discussion, here is a simple
strawman categorization of current work group efforts in HCLS SIG and
how conceivably might work together. This is not meant to prejudge
where each individual group might decide to go on its own. It's
simply my view of a way in which they might be synergistically aligned.
(1) Ontology Best Practices
Goal: Recommend and explain a set of best practices & most useful
current approaches for conducting our ontology-building activities.
Benefit: Educate the work groups, develop consensus on approaches,
support useful conventions, increase productivity of efforts
(2) Knowledge Lifecycle
Goals:
Develop an ontology & pilot applications characterizing at a high
level, digital resources in the lifecycle of scientific knowledge
production and consumption, such as publications, experimental data,
manuscripts, etc, their relationships to each other and to domain
ontologies in general. This effort would be mainly about "containers"
of knowledge, and their interrelationships and transitions, rather
than characterizing the nature of the content, which would be focused
on in Bio-RDF.
Benefit: Provide a common interoperability layer between existing
resources and multiple domain ontologies and resources.
(3) Bio-RDF
Goals:
(a) Develop and/or adopt useful RDF Schema representations for key
domain ontologies and "catalog"-type resources (ie.g., GO, MeSH,
UniProt etc) & build pilot applications which extract structured
content from current resources and convert into RDF
(b) Assess and test technological infrastructure for supporting RDF
representations, make recommendations and identify areas where
further work may be needed.
Benefit: Formalize existing resource annotation and catalog resources
sources and make them useful as real RDF content. Provide
information to the HCLS as a whole about technological approaches and
their tradeoffs.
(4) Structured/Unstructured
Goal: Identify and experiment in resolving any issues specific to
extraction of knowledge from text
Benefit: Provide a common experience base in these approaches and
identify key impacts and needs in this area.
(5) PROCESS MODEL
Goal: Develop and/or recommend semantic web compatible approaches to
modelling formal workflow including state machines, and
characterizing data provenance in automated environments
Benefits: Support the integration of the extremely common pattern of
computational pipelines of various types, into the semantic web
paradigm, integrated with the work of the other four groups.
Best
Tim
------------------------------------------------------------------------
------
Tim Clark
Director of Informatics
MassGeneral Institute for Neurodegenerative Disease
114 16th Street / CNY 114-2012, Charlestown, MA 02129
617-947-7098 (mobile) 617-724-1480 (fax)
www.mindinformatics.org
------------------------------------------------------------------------
------