My current research
I am currently developing machine learning methods for pharmacogenomic prediction of important pharmacogenes and pharmacodrugs from numerous data sources, mainly focusing on text mining. More information coming soon.
Projects I have worked on in the past
Over the past several years I have worked on a number of projects:
(1) Pharmspresso: a tool for semantic search in full-text articles in which concepts of interest to pharmacogenomics researchers are automatically identified, and the corpus may be searched for facts and relationships between concepts. See http://pharmspresso.stanford.edu/ (paper submitted, under review)
(2) the Health e-Decision project: a knowledge-based method for building patient decision-analytic tools, in which we developed software methods to deliver health decision aids through patient Web portals for electronic health records (read paper)
(3) an information extraction project in which we extracted subject demographic information from abstracts of all published randomized clinical trial reports, so as to allow consumers to search for trials and treatments relevant to their specific diseases and demographic group (read paper)
(4) LesionViewer: a tool for tracking cancer lesions which facilitates communication between medical personnel by data reduction techniques and subsequent employment of a rule-based system and temporal abstraction to make predictions about patient disease (paper to be published October 2007)
In the more distant past...
I worked on computational prediction of transcriptional regulatory network architecture from genome-wide DNA-protein interaction data (read paper).