Benchmarking accelerated next-generation sequencing analysis pipelines

Our Accelerated NGS analysis benchmark paper

• Link to the paper

Hardware Platforms for NGS analysis

Hardware accelerated platforms

Significant performance gain can be achieved with any accelerated platform

Runtime comparison

Above is a cartoon figure to highlight the patters the original paper contains correct figures

Speedup ## Quantify performance improvements (Speedup = CPU-only runtime ÷ Accelerated runtime)

Variant calling showed higher speedup than Mapping

Speedup comparison

Above is a cartoon figure to highlight the patters the original paper contains correct figures

Parabricks resource usage provided new insights

Profiling Parabricks

Above is a cartoon figure to highlight the patters the original paper contains correct figures

Optimizing Parabricks for further performance gains

Solution

Above is a cartoon figure to highlight the patters the original paper contains correct figures

Parabricks and DRAGEN, speedup and cost consideration

Practical Selection Framework

Above is a cartoon figure to highlight the patters the original paper contains correct figures

Accelerated NGS analysis NextFlow pipeline