20 years of statistical bioinformatics: a brief history of the limma and edgeR packages

Join the Vic branch for our post-AGM talk as we hear from Professor Gordon Smyth

Abstract

Genomic technologies underpin modern biomedical research. These technologies produce data that is hugely multidimensional. Statistical thinking is highly relevant but classic univariate statistical methods can perform poorly in the high dimensional context. The limma and edgeR software packages were developed for analysing gene expression data from microarray and RNA-seq technologies. This talk will give a brief historical perspective on the packages and a discussion of some of the ideas that made them successful.

Biography

Gordon has a PhD in mathematical statistics from the Australian National University. For 15 years, he was an academic in mathematics and statistics departments at the University of California, the University of Southern Denmark, the US Naval Postgraduate School and the University of Queensland. His research interests were in statistical modelling and statistical computing and he wrote a textbook on generalized linear models. Excited by the possibilities of the genomic era, he made the decision to move to the WEHI in 2001 to apply his skills to the new discipline of statistical bioinformatics. At the WEHI, Gordon pioneered the use of linear models, empirical Bayes and generalized linear models for modelling gene expression data from high-throughput genomic technologies. He and his Lab created what are still the world's most downloaded R packages for microarray and RNA-seq data analysis. He is currently Head of WEHI's Bioinformatics Division.