Using the Power of Modern Processors in Bioformatics

Research output: Contribution to conferencePosterResearch

Documents

Bioinformatics focuses on developing computational methods for
collecting, handling and analyzing biological data. Because the amount of
data is often very large and the models used for analysis are complex,
the demand for efficient methods exploiting modern computer hardware is
increasing.
In the past ten years we have been moving from single core to multi-core processor
architectures, and recently Graphics Processing Units (GPUs) with hundreds of cores
have also become available for general purpose computation. Making existing and
new algorithms exploit these multi-core architectures is one of today’s major
challenges in e.g. bioinformatics. At the Bioinformatics Research Center (BiRC), we
are currently working on two projects involving the exploitation of modern multi-core
architectures:
1. In molecular docking we aim to identify small proteins called ligands which show
strong interaction with a target protein molecule. Such ligands potentially alters
the function of the (disease causing) target protein and can therefore be used in a
modern drug discovery process. Evaluation of ligands against a target molecule
requires a moderate amount of computation, but we often need to evaluate
thousands of ligands against several target molecules, requiring use of expensive
computer clusters. By using GPUs to exploit the parallel nature of molecular
docking, we reduce the hardware requirements of large scale molecular docking
significantly.
2. Hidden Markov Models (HMMs) is a statistical tool used in a wide range of
applications within bioinformatics for modeling sequential data assumed to
originate from a Markov process; e.g. gene annotation, alignments and inferring
coalescence processes among species. Because of their computational efficiency,
HMMs are one of few methods used for genome wide analysis, where sequences
often consist of millions of characters. Nevertheless analysis times are still often
measured in days, weeks or months, and the models must therefore be kept
relatively simple. By exploiting SSE instructions and the multi-core architecture
of modern CPUs we decrease the running time of the analyses significantly and
thereby make analyses of more complex models feasible.
Original languageEnglish
Publication yearJun 2010
Number of pages1
Publication statusPublished - Jun 2010
EventConference on IT research at AU - Århus, Denmark
Duration: 3 Jun 20103 Jun 2010

Conference

ConferenceConference on IT research at AU
CountryDenmark
CityÅrhus
Period03/06/201003/06/2010

See relations at Aarhus University Citationformats

Download statistics

No data available

ID: 33899470