Quantifying the association between multidimensional phenotypes and multidimensional genetic/genomic data

General info

Date from - to

01 Sep 2005 - 31 Dec 2009

Project leader(s)

Prof. dr. A.H. (Koos) Zwinderman

Participant(s)

Menezes, de Renee Dr.

Leunissen, Jack Prof. dr.

ter Braak, Cajo Prof. Dr.

Abstract

In this project bioinformatics tools are developed to (quantitatively) study the association between phenotypic variables, and genetic, genomic, or proteomic measurements. The various sub-projects concentrate on the generalisation of existing multivariate statistical techniques to handle data with different measurement scales; modelling the joint distribution through a chain of conditional distributions; the overfit-problem – having many more variables than individuals, which results in a large number of pseudo-relations; and finally, non-parametric smoothing and data visualisation by exploiting the intercorrelations between neighbouring genes.

Link to the end report of this project

Publications

• Regression by L1 regularization of smart contrasts and sums (ROSCAS) beats PLS and elastic net in latent variable model
• Multivariate analysis of complex gene expression and clinical phenotypes with genetic marker data
• Penalized canonical correlation analysis to quantify the association between gene expression and DNA markers
• Integrated analysis of DNA copy number and gene expression microarray data using gene sets
• Can subtle changes in gene expression be consistently detected with different microarray platforms?