To view the poster abstracts, click here. Computational identification of microRNAs and their targets Dr. Volker Brendel miRNAs (microRNAs) are small non-coding RNAs that regulate mRNA degradation through complementary base pairing with specific target mRNAs. In recent years, more than 300 miRNAs have been identified in animals and plants using a combination of cloning and computational approaches. Using comparative genomics, it should be possible to identify the complete set of miRNAs in a genome. I will review several such studies involving RNA fold prediction and sequence property and similarity measures and report on our own results predicting miRNA genes in Arabidopsis thaliana and rice. Pseudogenes and Gene Conversion Dr. Todd E. Scheetz Pseudogenes are prevalent throughout the human genome, as well as most higher eukaryotes. They are duplications of existing genes that have accumulated disabling mutations due to lack of evolutionary pressure to remain functional. They are thought to play a role in evolution, providing a source from which sequence can be "borrowed" to build a new gene or alter an existing gene's function. However a similar process, gene conversion, is known to play a role in the development of several diseases. A previous study performed at the EMBL identified over 19,000 pseudogenes. Our own investigation has identified additional pseudogenes within the human genome. Annotation and prioritization of these pseudogenes is currently underway in our lab, specifically focusing on the issues of expression and gene conversion. Metabolic Flux Maps of Central Carbon Metabolism in Plants Dr. Jacqueline V. Shanks Metabolic flux quantification is important in the detailed understanding of metabolism. Since fluxes provide a quantitative depiction of carbon flow through competing metabolic pathways, they are an important physiological characteristic akin to levels of transcripts, proteins, and metabolites. However, flux measurements are nontrivial to obtain on a systematic level, especially in plants. We have developed a tool, NMR2Flux, for metabolic flux analysis of soybean embryos and C. roseus root tissues. This tool combines NMR analysis of biosynthetically derived fractional 13C labeling of proteinogenic amino acids and starch with metabolite balance models. It provides solutions for carbon fluxes through primary metabolic pathways in the cytoplasm, mitochondria and plastids essential for protein, oil, and starch synthesis. The technology is extremely powerful because it considers all of intermediary metabolism - glycolysis, pyruvate metabolism, pentose phosphate pathway, tricarboxylic acid sycle and C1 metabolism. Phylogenomic analyses of the origins and evolution of meiosis Dr. John Logsdon Our ongoing bioinformatic inventory of meiotic genes in eukaryotic genomes indicates that numerous genes-many of which are meiosis-specific-are found among diverse eukaryotic lineages. Our analyses include a variety of protists, including some species thought to be both early-diverging and asexual. Giardia, for example, contains at least five meiosis-specific genes in addition to many other genes implicated in both meiosis and DNA repair. The evolutionary relationships of these genes have been validated with rigorous phylogenetic analyses. These results suggest that meiosis evolved early in eukaryotic evolution and that organisms previously thought to be primitively asexual may be capable of meiotic sexual reproduction. Virus dynamics and evolution during persistent lentivirus infection Dr. Susan Carpenter Lentiviruses exist in vivo as a population of related, non-identical genotypes, commonly referred to as a quasispecies. The quasispecies structure is characteristic of complex adaptive systems and contributes to the high rate of evolution in lentiviruses that confounds efforts to develop effective vaccines and antiviral therapies. Experimental infection of horses with equine infectious anemia virus (EIAV) provides a unique system for longitudinal studies of lentivirus-host interactions during stages of clinical disease and clinical quiescence. We will discuss how computational approaches have aided our understanding of EIAV dynamics and evolution in vivo. Selection of 5' ESTs for full-insert sequencing Dr. Thomas Bair We have developed a system for prioritizing the selection of full-lengt h inserts from 5 prime-end EST sequenced cDNA libraries. The selected c lones are full-insert sequenced and assembled to high quality standards -- an expensive process that increases the importance of selecting a m inimal number of clone candidates, which do not include the true 5' end of the native transcript. It is also important however, to include as much diversity in the selection as possible. Our clones, when selected, are deposited into the Mammalian Gene Collection (MGC) -- a collection of over 13,000 freely available mouse clones (http:// mgc.nci.nih.gov/ ). To date we have submitted 1585 clones to the MGC project, while this is a small fraction of the total, our submissions comprise the majorit y of the longer transcripts with transcripts in the 5-7 kb size range c oming mostly from our group. Our system has evolved from initially finding clones homologous to know n and predicted genes and proteins, to consideration of intrinsic chara cteristics of 5' EST sequences, such as the start codon, quality of a K ozak consensus sequence, length of the open reading frame etc., to our latest system that considers these features in addition to the genomic context. For example, our selections are enhanced by observing other ESTs and gene predictions in the same genomic location, as well as using finished genomic sequence rather than the typically more error filled ESTs. These details are fed into a decision tree machine learning system to pick clones, which then enter our full-insert sequencing pipeline. We have analyzed the success and utility of the various methods and will present a summary of effective metrics as well as a comparison of "genomic" clustering vs. a more traditional EST similarity-based approach (UI-cluster) and analyze trends in the two methods of clustering