Affymetrix gene-chips are now a standard tool for systematically interrogating genome-wide gene expression levels to understand the genetic contributions to a range of diseases. Although the information these assays provide is highly valuable, the platform suffers from an out-dated annotation system that cannot help interpret the data to its fullest potential.
Renaldo and his Broad colleagues are remapping the individual measurement features on the gene chips, taking into account up to date sequence information from a variety of transcript model databases. They will then validate that these redefined sets provide more reliable information regarding expression of the corresponding genes, by demonstrating increased consistency of associations between gene-expression levels and clinical features of analogous datasets.
These results (and the improved annotation tools) will be made freely available to the entire research community.
PROJECT: Ambiguous Probes Create Artificially High Correlation of Expression levels between Affymetrix Probe-sets
In my research project this summer I wrote several computer programs to determine if there were problems in the design of the Affymetrix GeneChip. I was very nervous when I first began my project because I only had a few weeks to do a very challenging project. I was struggling to teach myself two new computer languages and learn the basics about the technology I was studying, but I quickly found out that in the Broad collaborative environment I could find all the help I needed. Within a few weeks I felt like an expert in my subject talking to graduate students and PhDs, and when I finally presented my work I amazed myself with all I was able to accomplish in nine weeks.