Skip to main content

News & Events

CIRC Symposium Series

November 21, 2014
11:30 AM - 01:00 PM
Computer Studies Building (CSB) 209, River Campus

Cosmos in Computers: How Exascale Computing and Laboratory Experiments Improve our Understanding of the Universe

Pierre-Alexandre Gourdain, Ph.D., Department of Physics and Astronomy

Understanding our universe is a slow and expensive process. In recent years, astrophysicists have accumulated an enormous amount of observational data. Yet this process is passive (observations rather than experiments) and takes place on universal time scales (years rather than days). This talk will show how experiments done in the laboratory can improve the numerical simulations of astrophysical phenomena such as plasma jets generated by pre-planetary nebulae and active galactic nuclei. In turn numerical simulations can be scaled to astrophysical scale to verify the accuracy of physical models.

On-Going Research Talk: miRge: A Rational, Ultrafast, Multiplexed Method of Processing miRNA RNA-seq Data

Jason Myers, University of Rochester Genomics Research Center

Small RNA RNA-seq for microRNAs (miRNAs) is a rapidly developing field in need of improved bioinformatics tools. Current analysis methods have been modeled on successful methods of RNA-seq analysis for mRNA or genomic DNA projects but are not optimized for miRNAs. We set about to develop a rational, ultrafast and multiplexed method to process miRNA RNAseq data. We built miRge, a Perl-based tool which utilizes a hierarchical multistep processing of fastq-formatted small RNAseq data taking advantage of previously built tools, novel library modifications and aligners. Each fastq file, representing a sample that is being analyzed by RNA-seq for miRNAs, is first quantitated by removing linker sequences and tabulating all unique nucleotide sequences identified, ultimately resulting in a table of counts for all unique nucleotide sequences across all samples. The complete set of unique nucleotide sequences identified in this table is then annotated via a sequential alignment process against customized mature miRNA, hairpin miRNA, noncoding RNA and mRNA sequence libraries using bowtie alignment software. Output is given for miRNAs as raw reads and reads per million (RPM) in addition to isomiR data and reads for other RNA species. miRge processed a 2.4 million read RNA-seq file in 24 seconds and a 33 million read file in 2.5 minutes. In comparisons with miRExpress 2.0, sRNAbench, omiRAs and miRDeep2, miRge was consistently faster and generally aligned more miRNA reads per sample. miRge is a rational, efficient, and modular approach to miRNA RNA-seq data processing. It provides a significant improvement in processing speed and quality of alignment to known miRNAs. miRge is available at http://atlas.pathology.jhu.edu/baras/miRge.html

Category: Talks