FPGA Implementation

Facts & Details: 

Many research projects suffer from long computation times of bioinformatics applications. A reconfigurable computing architecture based on Field-Programmable Gate Arrays (FPGAs) has the advantage of being several magnitudes faster than a regular large high-performance compute cluster (HPC), while at the same time it is significantly smaller in size (comparable to a regular desktop computer) and consumes a lot less energy. FPGAs are commonly used in cryptography where extensive repetitive computations are necessary.

In a joint undertaking with the Department of Technical Computer Science in Kiel, we have set up an FPGA instrument, the RIVYERA FPGA deviceS6-LX150, which provides ultra-fast, ready-to-use bioinformatics applications: for example, Smith-Waterman alignment, BLASTN and BLASTP run on the system. Our current work focuses on the acceleration of state-of-the-art bioinformatics applications, e.g. interaction analyses (epistasis), short-read alignment, exact matching for HLA alignments and imputation of genotyping or sequencing data.

Figures 2A and 2B show how RIVYERA accelerates Smith-Waterman exact alignments by orders of magnitude compared to other computing platforms while saving more than 99% of energy (“green IT”).Our work further serves as an interdisciplinary research project between design artists and researchers. Together with experts for graphical user interface design from the Muthesius Academy of Fine Arts and Design in Kiel we are developing a user-friendly graphical front-end for the RIVYERA system.

                           Model of a user-friendly graphical interface for exact HLA sequence alignments.

 

 

 

 

 

 

 

In our collaboration, Prof. Jacob and Benjamin Schwarz, a former Master student of Prof. Jacob, designed and finalised a graphical user web-interface prototype to enable quick and easy phasing and imputation from diverse sets of worldwide reference populations. The “conversational interface” developed by Prof. Jacob and Benjamin Schwarz allows the selection of reference samples from a diverse set of worldwide reference populations (Figure 4). For more information incl. a short video see here and the example in the picture below..

conversational interface - example web

 

Collaboration partners:
Prof. Dr. rer. nat. Manfred Schimmler, Lars Wienbrandt and Jan Christian Kässens, Kiel University
Prof. Dr. Frank Jacob and Simon Koch; Muthesius Academy of Fine Arts and Design in Kiel

Link to a FPGA company in Kiel: http://www.sciengines.com

Main Publications

show all publications