Our Projects
Here we list some current projects currently underway. We are seeking funding from a number of sources to expand and elaborate on basic elements described here.
'Snapshots of a subject'
The annual meeting for the Society for Neuroscience is attended every year by a large proportion of the world's neuroscientists. Every year over 10,000 abstracts are published by the Society in their support material for the meeting. We use topic modeling to analyze the contents of this collection over the period 2000-2007 and then use this data to construct a navigable map of the subject of neuroscience. We performed this analysis for the abstracts of the 2006 annual meeting and to create a navigable map (accessible here). This work was performed in collaboration with David Newman and Padhraic Smyth at UC Irvine (who performed the topic modeling analysis) and Bruce Herr at the Cyberinfrastructure for Network Science Center at Indiana University (who performed the visualization work).
'Accelerating Biocuration'
As described in the 'home' page, one of the key vulnerabilities of modern bioinformatics is speed at which biocurators can extract information from the text of articles to be entered into biomedical databases. We are developing information extraction technology (based on well-established computer science principles and the latest technology) to construct a framework to accelerate biocuration. We have published several conference papers and and gave an invited presentation at the 2nd annual International Biocurator's meeting (presentation available here) concerning this. We are actively collaborative projects with Anita de Waard at the University of Utrecht and Elsevier to mine neuroscientific articles for mentions of brain regions. We are pursuing funding to collaborate with the NCBO and the Jackson Laboratory to accelerate the biocuration of phenotype information for mice.
BioScholar
Previous development work included the NeuroScholar knowledge management system and NeuARt II neuroanatomical mapping application. The next iteration of these systems ('BioScholar')?will permit the head of a research group to set up and manage a small scale informatics system describing the research literature pertaining to his/her subject. We intend to make this possible based on (50%) participation of a local biocurator at the level of a senior graduate student or postdoc. We are seeking RO1-level funding from NIH to accomplish this and our first application for this was scored in the 17.1th percentile.
We are developing a knowledge representation and reasoning strategy by concentrating on the relationships between variables that are constrained and those variables that are measured in experiments. This work is still at an early conceptual stage and has similarities to the work of Ida Sim on computaional approaches to clinical trials (TrialBank). We are actively participating within the ontology engineering community (specifically development work within the Ontology for Biomedical Investigation, OBI) to integrate our work directly into their wider effort. This idea provides a viable low level framework for our entire effort and will support innovative future work to develop the next-generation of biomedical informatics infrastructure.
Collaborator Links
'Snapshots of a subject'
Places and Spaces: Mapping Science
'Accelerating Biocuration'
The National Center for Biomedical Ontology
BioScholar
Knowledge Engineering from Experimental Design (KE-f-ED)
