High-Dimensional Sparse Factor Modeling: Applications in Gene Expression Genomics

By Carlos M. Carvalho, Jeffrey Chang, Joseph E. Lucas, Joseph R. Nevins, Quanli Wang, M. West. Journal of the American Statistical Association, 2008

574 citations recorded by the source corpus.

We describe studies in molecular profiling and biological pathway analysis that use sparse latent factor and regression models for microarray gene expression data. We discuss breast cancer applications and key aspects of the modeling and computational methodology. Our case studies aim to investigate and characterize heterogeneity of structure related to specific oncogenic pathways, as well as links between aggregate patterns in gene expression profiles and clinical biomarkers. Based on the metaphor of statistically derived “factors” as representing biological “subpathway” structure, we explore the decomposition of fitted sparse factor models into pathway subcomponents and investigate how these components overlay multiple aspects of known biological activity. Our methodology is based on sparsity modeling of multivariate regression, ANOVA, and latent factor models, as well as a class of models that combines all components. Hierarchical sparsity priors address questions of dimension reduction and multiple comparisons, as well as scalability of the methodology. The models include practically relevant non-Gaussian/nonparametric components for latent structure, underlying often quite complex non-Gaussianity in multivariate expression patterns. Model search and fitting are addressed through stochastic simulation and evolutionary stochastic search methods that are exemplified in the oncogenic pathway studies. Supplementary supporting material provides more details of the applications, as well as examples of the use of freely available software tools for implementing the methodology.

Subfields: Molecular Biology

View Paper: https://www.semanticscholar.org/paper/c41783252096992982999e543b86e6d7c8af9c6d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

Symmetric Variational Autoencoder and Connections to Adversarial Learning (2017, background): “, factor scores) that are inferred given observed data; the latent variables are often used for a down-stream goal, such as classification [Carvalho et al., 2008].”
Testing for Associations between Loci and Environmental Gradients Using Latent Factor Mixed Models (2012, methodology): “For each series of experiments, we generated 10 replicates of this generative model, and we studied the distributions of P -values for tests using LFMMs.”
Should Regularization Replace Simple Structure Rotation in Exploratory Factor Analysis? (2019, background): “This is especially important for data sets with a very large number of variables (relative to the sample size) such as genome data (e.g., Carvalho et al., 2008), or fMRI data (e.g., Hirose, 2016).”
Large Covariance Estimation by Thresholding Principal Orthogonal Complements (2011, unknown): “The spectral consistency also enhances the results in Chamberlain and Rothschild (1983).”
Enhancing Interpretability in Factor Analysis by Means of Mathematical Optimization (2019, unknown): “…Fontanella, & Adachi, 2017; Witten et al., 2009; Yamamoto et al., 2017), Bayesian frameworks, (Kaufmann & Schumacher, 2017; Zhao, Gao, Mukherjee, & Engelhardt, 2016), clustering approaches, (Adachi & Trendafilov, 2018; Uno, Satomura, & Adachi, 2016) or regression models, (Carvalho et al., 2008).”
Bayesian group factor analysis with structured sparsity (2016, methodology): “One traditional solution is to restrict the loading matrix to be lower triangular (West, 2003; Carvalho et al., 2008).”
LFMM 2: Fast and Accurate Inference of Gene-Environment Associations in Genome-Wide Studies (2019, methodology): “Here, we extend the definition of the LFMM data matrix to DNA methylation profiles and other molecular markers within a unified framework (Leek and Storey 2007; Carvalho et al. 2008).”
Spatial dynamic factor analysis (2008, background): “(b) y t processes at t = 10, 11, 12. vated by recent interest in the factor models literature towards hierarchically structured loadings matrices (see Lopes and Migon, 2002, Lopes and Carvalho, 2007, Carvalho et al. 2008, and West, 2003, to name a few).”
Exploiting Structure in Wavelet-Based Bayesian Compressive Sensing (2009, unknown): “resentation of the full posterior [35], [39]).”