Integrative analysis of multiple cancer prognosis studies with gene expression measurements

Abstract

Although in cancer research microarray gene profiling studies have been successful in identifying genetic variants predisposing to the development and progression of cancer, the identified markers from analysis of single datasets often suffer low reproducibility. Among multiple possible causes, the most important one is the small sample size hence the lack of power of single studies. Integrative analysis jointly considers multiple heterogeneous studies, has a significantly larger sample size, and can improve reproducibility. In this article, we focus on cancer prognosis studies, where the response variables are progression-free, overall, or other types of survival. A group minimax concave penalty (GMCP) penalized integrative analysis approach is proposed for analyzing multiple heterogeneous cancer prognosis studies with microarray gene expression measurements. An efficient group coordinate descent algorithm is developed. The GMCP can automatically accommodate the heterogeneity across multiple datasets, and the identified markers have consistent effects across multiple studies. Simulation studies show that the GMCP provides significantly improved selection results as compared with the existing meta-analysis approaches, intensity approaches, and group Lasso penalized integrative analysis. We apply the GMCP to four microarray studies and identify genes associated with the prognosis of breast cancer.

Promoting Similarity of Sparsity Structures in Integrative Analysis With Penalization

“…Under the sparsity condition, many MTL studies have adopted the group Lasso to achieve a common set of features for multiple datasets (Argyriou et al, 2008; Lounici et al, 2009; Liu et al, 2009; Yuan et al, 2012). As the MCP-based group selection outperforms the Lasso-based (Ma et al, 2011), we adopt the group MCP for MTL analysis. The third method for comparison is the composite MCP, which can provide direct insights into the newly developed penalty.…”

Sparse group penalized integrative analysis of multiple cancer prognosis datasets

“…In Huang et al (2012c) and Ma et al (2011a, 2011b), the homogeneity model is adopted to describe the genomic basis of M datasets. We briefly describe this model here for integrity of this article.…”

“…Recent studies have shown that pooling and analyzing multiple studies may effectively increase sample size and improve properties of the identified markers (Guerra and Goldsterin 2009). For example, simulation and data analysis have shown that markers identified in multi-dataset analysis may have more true positives, fewer false positives, and better prediction performance (Ma, Huang and Moran 2009; Ma, Huang and Song 2011a; Ma et al 2011b). Multi-dataset methods include meta-analysis and integrative analysis methods.…”

“…In studies such as Ma et al (2011b), the homogeneity model has been assumed. Under this model, multiple datasets have exactly the same set of markers.…”

“…Such differences may make the homogeneity assumption too restricted. In addition, data analyses in Ma et al (2011a, 2011b) show that for some identified genes, the magnitudes of estimated regression coefficients may vary significantly across datasets. It is possible that the very small regression coefficients are actually zero.…”