FEATURED SPEAKER: Claus Kadelka, Assistant Professor in the Department of Mathematics
TITLE: Concise functional enrichment of ranked gene lists
Background: The integration of large scale biology and technology has led to an explosion of genome-wide expression data. Given a set of differentially-expressed genes as input, functional enrichment methods compute a set of functional categories that annotate a surprisingly large number of the genes. Many functional enrichment methods work in a binary mode, i.e., they consider a gene to be either expressed or not. These approaches may disregard useful information provided by the ranking. The few enrichment methods that do operate on ranked lists output highly redundant or non-specific functional categories.
Results: We present a new functional enrichment algorithm, called Concise Ranked Functional Enrichment (CRFE), that effectively utilizes the ranking in gene expression data sets and computes a small set of non-redundant functional categories that are significantly over-represented in highly ranked genes. We apply CRFE and four existing functional enrichment methods to two treatment-control microarray data sets and compare their performance based on utilization of ranking information, conciseness and specificity. CRFE performs well in all criteria, whereas every other method performs poorly in at least one. Lastly, we report a high-level interpretation of the functional categories returned by CRFE for each data set.
Conclusion: We introduce a novel functional enrichment algorithm that uses the ranking of genes in a list to compute functional categories that annotate many of the highly ranked genes, and compare its performance with four other methods.