Gene set (pathway and network) based genomic data analysis

• Acknowledgements: This study has been supported by
  • R03LM009754 from NLM/NIH (National Library of Medicine): Effective clustering penalized methods for genomic biomarker selection. Funding period: 08/01/2009--07/31/2010. PI: Shuangge Ma.
  • DMS-0904181 from NSF (DMS): Novel methods for pharmacogenomic data analysis using gene clusters. Funding period: 8/15/09--8/14/12. PI: Shuangge Ma. Co-PI: Michael Kosorok, Department of Biostatistics, UNC Chapel Hill.
  We would like to thank members of Yale Cancer Center for insightful discussions.
  
  
• Goal: The goal of this study is to develop statistical methods that can make more effecient use of cancer genomic data by properly accounting for the clustering (pathway or network) structure of gene expressions and selecting predictive cancer biomarkers. It can provide more insights into the genomic mechanisms of cancer occurrence and progression.
  
  
• Publications:
  1. Huang, J., Ma, S., Xie, H. and Zhang, C. (2009) A group bridge approach for variable selection. Biometrika, 96(2), 339-355.
  2. Ma, S., Zhang, Y., Huang, J., Han, X., Holford, T., Lan, Q., Rothman, N., Boyle, P. and Zheng, T. (2010) Identification of Non-Hodgin's lymphoma prognosis signatures using the CTGDR method. Bioinformatics, 26, 15-21.
  3. Ma, S. and Kosorok, M.R. (2010) Detection of gene pathways with predictive power for breast cancer prognosis. BMC Bioinformatics, 11:1.
  4. Ma, S., Huang, J., Shi, M., Li, Y. and Shia, B. Semiparametric prognosis models in genomic studies. Briefings in Bioinformatics. In Press.
  
  
• Posters:
  1. Gene network analysis and identification of lymphoma prognosis markers. 2010 Clinical and Translational Research and Education Meeting ACRT/SCTS Joint Annual Meeting.
  
  
• Oral Presentations:
  1. Identification of cancer-associated gene pathways from analysis of expression data. JSM. August 3rd, 2009.
  2. Variable selection in the accelerated failure time model via the bridge method. IMS, China. July 3rd, 2009.
  3. A Tale of Two Streets: Incorporating grouping structure in high dimensional data mining. Data Mining and Business Intelligence Conference. June 6th, 2009.
  4. Two-level gene selection via group bridge penalization. Department of Statistics, Columbia University. Jan. 22nd, 2009.
  
  
• Please email shuangge.ma AT yale.edu for more information.