DNA assembly with de bruijn graphs on FPGA.

Area of Science:

• Computational Biology
• Bioinformatics
• Computer Engineering

Background:

• DNA assembly is a critical step in synthetic biology and genomics.
• Traditional computational approaches for DNA assembly can be resource-intensive.
• Field-programmable gate arrays (FPGAs) offer potential for hardware acceleration.

Purpose of the Study:

• To evaluate the effectiveness of FPGA-based accelerators for the Ray DNA assembly algorithm.
• To compare the performance and energy consumption of an FPGA implementation against a CPU-based approach.

Main Methods:

• The Ray algorithm was reprogrammed using the OpenCL language for parallelization.
• Optimizations were implemented to adapt the algorithm for FPGA architecture.
• Performance and energy consumption were benchmarked on representative datasets.

Main Results:

• The FPGA-accelerated Ray algorithm demonstrated superior performance compared to the CPU version.
• Significant reductions in energy consumption were observed with the FPGA implementation.
• FPGA platforms are shown to be a viable and efficient solution for DNA assembly.

Conclusions:

• FPGA accelerators provide a powerful and energy-efficient alternative for DNA assembly.
• The optimized Ray algorithm on FPGAs significantly outperforms traditional CPU-based methods.
• This work validates the use of FPGAs for accelerating complex bioinformatics tasks.