Bioinformatics is an interdisciplinary research area that applies computer and information science to solve biological problems. On February 8, 2002, the Graduate Program in Bioinformatics (http://www.binf.umn.edu/) was approved by the University of Minnesota Regents. This program now offers a minor in bioinformatics to students enrolled in a UM graduate program. We will describe the present status of and future plans for the graduate program.

Regulation of gene expression at the level of mRNA decay facilitates rapid, selective and temporally precise responses to T cell activation. We have used oligonucleotide microarrays to profile the decay rates of mRNA transcripts in resting and stimulated T cells. Data from four independent experiments was analyzed using a first order decay model and RNA half-lives (with 95% confidence intervals) were determined for each of approximately 6000 expressed transcripts. While the majority of transcripts in resting T cells were stable, our data identified numerous short-lived transcripts encoding important regulatory genes including transcription factors, signal transduction regulators, cell cycle regulators, and regulators of apoptosis. In addition, we found that T cell activation led to dramatic and statistically significant changes in the decay rates of numerous transcripts; in some cases the mRNA was stabilized and in others it was destabilized. The technology employed in this study allowed the identification of numerous genes that are regulated at the level of mRNA stability. Our data demonstrate that the rate of mRNA decay of numerous important regulatory genes changes in a stimulus-dependent manner, highlighting the importance of mRNA stability in the regulation of gene expression.

Superfamilies are composed of distantly related proteins that share functional or structural connections to a common ancestor. Low sequence identity causes difficulties in identifying additional superfamily members. In this study, a combination of PSI-BLAST and SHOTGUN, a program that correlates the output from thousands of individual PSI-BLAST analyses, was used to expand the amidohydrolase superfamily, with a low false positive rate. The amidohydrolase superfamily contains diverse hydrolytic enzymes that utilize divalent metals to activate water for nucleophilic attack on the substrate. High conservation of three dimensional structure and catalytic mechanism across this superfamily has been observed in members where crystal structures are available, such as urease, adenosine deaminase, and phosphotriesterase. Among the reactions catalyzed by superfamily members are deamination, deamidation, dehalogenation, and dephophorylation. Further characterization of this superfamily in sequence space could provide insights into how evolution expands enzymatic functionality without altering or having to engineer new protein scaffoldings.

The Supercomputing Institute (www.msi.umn.edu) was created in 1984 to provide state-of-the-art high-performance computing resources to the University of Minnesota research community. The Institute has built a strong tradition of providing researchers with not only leading edge hardware and software, but also providing technical user support in the form of tutorials, workshops, and one-on-one assistance. Current technical user support staffs have expertise in bioinformatics, molecular modeling and drug design, computational chemistry, scientific visualization, computational fluid dynamics, structure mechanics, and geophysics.

The Supercomputing Institute is committed to providing the computational resources (hardware, software, technical assitance, etc.) that are required to keep the University of Minnesota computational biology research community on the leading edge of academic research. Please stop by and talk with us at our poster to discover how the Institute can help make your research more efficient. Contact us by phone, 612-626-0802, or by email, help@msi.umn.edu if you have questions before or after the symposium.

Sequence based homology search is the workhorse of bioinformatic analysis. The algorithmic mechanics underlying homology search tools are poorly understood by the major part of the user community. Comparative studies are rare, ad-hoc, and usually performed as part of the testing of a proposed new algorithm. Such new algorithms would never be written except in response to perceived flaws in existing tools, so the authors are hard pressed to maintain impartiality. We present several sets of test data representing divergent real-world use cases for sequence based homology search algorithms. We define several metrics for comparison, and present visualizations of these metrics which allow quick, clear communication of the results. Example results are presented for several homology search tools.