Search

Normalization procedures attempt to remove non-biological variance found within micro array datasets. The choice of normalization procedure is important, as it has a dramatic effect on downstream data analysis. Although many normalization procedures have been developed, comparison and evaluation of their performance is difficult. We present a method to evaluate normalization procedures by utilizing gene-gene associations derived from the biomedical literature via Latent Semantic Indexing. The functional coherence of co-regulated genes obtained from different normalized data sets is calculated in order to evaluate the effectiveness of each normalization procedure. The method was tested on three popular normalization procedures (MAS5, PDNN and RMA) applied to gene expression across 71 recombinant inbred mouse brain samples. Results show that, on average, MAS5 outperforms both PDNN and RMA by producing a higher number of functionally cohesive gene sets. These results demonstrate that our literature-based cohesion analysis can provide an objective method for evaluation of normalization procedures.

Numerous algorithms exist for producing gene sets from high-throughput genomic and proteomic technologies. However, analysis of the functional significance of these groups of genes or proteins remains a big challenge. We developed a web based system called Gene-set cohesion analysis tool (GCAT) for estimating the significance level of the functional cohesion of a given gene set. The method utilizes Latent Semantic Indexing (LSI) derived gene-gene literature similarities to determine if the functional coherence of a gene set is statistically significant compared to that expected by chance. The robustness of the method was determined by evaluating the functional cohesion for over 6000 Gene Ontology categories. Here, we demonstrate the utility of GCAT for analysis of microarray data from previously published experiments in which embryonic fibroblasts were treated with interferon. Using GCAT, we found the highest literature cohesion p-value (p= 1.37E-63) corresponded to a set of genes that were differentially regulated > 2-fold and had a t-test p-value 2-fold (literature p-value=2.2E-44) or had a p-value 0.05 did not show a significant literature cohesion. These results demonstrate that GCAT can provide an objective literature-based measure to evaluate the biological significance of gene sets identified by different criterions. GCAT is available at http://motif.memphis.edu/gcat/.