Integrative Genomics and its Implications for Clinical Research and Care: What Are the Real Issues Beyond Analysis?

Monday, September 29, 2003 - 1:30pm - 2:20pm
Keller 3-180
Atul Butte (Children's Hospital Informatics Program)
Microarrays can provide systematic quantitative information on the expression of thousands of unique RNAs, and have been used to elucidate the transcriptional response in many basic biological and clinically relevant experiments, ranging from associative studies between therapeutics and expression, to aiding in diagnostic questions, to discovery of novel subtypes of disease.

Given the over 6000 arrays with data publicly available and the surfeit of microarray facilities, the rate-limiting steps is no longer the sample collection, hybridization, scanning, or even the analysis. Instead, the new challenge is in taking findings, such as the traditional list of genes resulting from a microarray analysis, and ascertaining the meaning of the results, such as the biological relationships between the genes. However, tools that link these genes back to known biological pathways, as well as discovering new pathways, are in their infancy. Tools that automatically suggest the importance of particular findings have yet to be invented.

During this presentation, I will describe four packages we have made freely available to the academic genomics community. I will present examples of and would like to discuss (1) not all pathways will be reverse engineered using microarrays, (2) looking for simultaneous gene associations ignores the fact that biology takes time, (3) a discovered diagnostic model doesn't imply the underlying molecular physiology, (4) due to rapidly changing information about the genes already measured, one is never truly finished analyzing a microarray dataset, and (5) the real bottleneck in microarray analysis is not the analysis, but the interpretation of the findings.