Sparse models for imaging genetics

摘要: Imaging genetics is an emerging and promising technique that investigates how genetic variations affect brain development, structure, function. Many imaging studies involve disorder-related neuroimaging phenotypes molecular data, such as single nucleotide polymorphism data. However, this class of challenging due to the relatively small number subjects but extremely high-dimensionality both In chapter, we introduce a suite sparse methods—that can produce interpretable models are robust overfitting—for genetics. Moreover, incorporate various biological prior knowledge—such linkage disequilibrium information—into analysis. nonsmooth highly complex regularizers, applications large-scale problems remain challenging. Thus novel optimization techniques, is, screening, which boost efficiency many on data sets by several orders magnitude.