ARMI: Assisted Robust Marker Identification
Hao Chai haochai2@gmail.com
Analysis of Cancer Gene Expression Data with an Assisted Robust Marker Identification Approach. Genetic Epidemiology, in revision.
In recent multidimensional studies, data have been collected on gene expressions as well as their regulators (for example, copy number alterations, methylation, and microRNAs), which can provide additional information on the associations between gene expressions and cancer outcomes.
The proposed ARMI approach borrows information from regulators and can be more effective than analyzing gene expression data alone. A robust objective function is adopted to accommodate long-tailed distributions and contamination. Marker identification is effectively realized using penalization.