Sparse Canonical Correlation Analysis

Sparse Canonical Correlation Analysis (SCCA)
performs data integration by simultaneous analysis
of 2 data types to find the relationships between
them. It is applicable to all studies including
large studies with limited sample size. SCCA
provides sparse solutions that include only a small
subset of variables. Sparse results aid in
biological interpretability and can be used for
hypothesis generation. This monograph presents
methodology for SCCA and evaluation of its
properties using simulated data. The practical use
of SCCA is illustrated by applying it to the study
of natural variation in human gene expression. Two
extensions of SCCA - adaptive SCCA and modified
adaptive SCCA are also presented. Their performance
is evaluated and compared using simulated data and
adaptive SCCA is applied to the study of natural
variation in human gene expression.

Autorentext

Dr. Elena Parkhomenko received a PhD in Biostatistics at the University of Toronto. Her research interests are in the area of data integration in the context of high-dimensional studies and discovery of modest effects in multiple-testing problems.

Klappentext

Sparse Canonical Correlation Analysis (SCCA) performs data integration by simultaneous analysis of 2 data types to find the relationships between them. It is applicable to all studies including large studies with limited sample size. SCCA provides sparse solutions that include only a small subset of variables. Sparse results aid in biological interpretability and can be used for hypothesis generation. This monograph presents methodology for SCCA and evaluation of its properties using simulated data. The practical use of SCCA is illustrated by applying it to the study of natural variation in human gene expression. Two extensions of SCCA - adaptive SCCA and modified adaptive SCCA are also presented. Their performance is evaluated and compared using simulated data and adaptive SCCA is applied to the study of natural variation in human gene expression.