Projects

Understanding asthma

• Molecular phenotyping of asthma:
  In a previous study, we identified 6 subject clusters from 378 SARP asthmatic and healthy control participants with bronchoscopic inflammatory and FENO data using unsupervised machine learning approaches. Our analysis reveals several clinically recognizable patient clusters. Using variable selection and supervised learning techniques, we also identified 51 relevant and nonredundant variables distinguishing subject clusters which include age of asthma onset, quality of life, symptoms, medications and health care utilization. These results suggest that machine learning-based analyses facilitate understanding of variable and subject clusters, as well as provide unique insights into asthma phenotyping, confirming other approaches while revealing novel additional phenotypes. Currently, we are investigating whether the subject clusters are reproducible in new clinical datasets.
• Understanding regulatory mechanisms of different subtypes of asthma:
  In this project, we are interested in addressing the following questions: what are the molecular mechanisms underlying different subtypes of asthma? Can we identify good drug targets for asthma treatment? Can we identify drugs for asthma treatment? We are collaborating with our collaborators at the University of Pittsburgh School of Pharmacy to try to find answers to these questions.

Understanding drug abuse

• NIDA Center of Excellence for Computational Drug Abuse Research (CDAR):
  This is a joint initiative between Dr. Eric Xing's and my labs at Carnegie Mellon University (CMU) and the University of Pittsburgh, funded by the NIH (NIDA). There are three research support Cores (Cores A, B and C) in the Center, and Core C - The Computational Genomics Core for DA (CG4DA) - is led by Dr. Xing (director) and me (co-director). The three themes of the services that CG4DA proposed to provide to the FRPs and also a broader DAR community are:
  • developing machine learning methods for transcriptome-wide screening of expression traits and molecular markers for DA;
  • genome-wide discovery of new drug targets and their epistatic genetic influences via structured association mapping;
  • software development for DA diagnosis, and towards guiding DA treatment.