Accelerating Minimap2 for Accurate Long Read Alignment on GPUs

Sadasivan, Harisankar; Maric, Milos; Dawson, Eric T.; Iyer, Vishanth; Israeli, Johnny; Narayanasamy, Satish

doi:10.26502/jbb.2642-91280067

Cited by 15 publications

(21 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Prior works [8,10,15] have found chaining to be the bottleneck in aligning long ONT reads with Minimap2 -40% to 70% of CPU time is spent in chaining. The chaining kernel's complexity is quadratic with respect to read length while 1 Intel is a trademark of Intel Corporation or its subsidiaries the seeding algorithm scales linearly.…”

Section: Quadratic Complexity Of Chainingmentioning

confidence: 99%

“…end for 11: end for 3.2 Split Score Generation Kernels: Specialize Kernels for Different Range Profiles In mm2-ax [15], Sadasivan et al report that the chaining kernel has highly irregular arithmetic intensity and intrarange parallelism, characterized by a high variance in the successor range. We extend this idea by measuring workload in terms of chaining scores computed to reflect the mapping cost more accurately.…”

Section: Algorithm 1 Segment Cutting In Range Selection Kernelmentioning

confidence: 99%

“…Our test data is extracted from Oxford Nanopore Technologies (ONT)'s 60X HG002 dataset [5]. To demonstrate our performance gain on different read lengths, especially long reads, we allocate five bins based on read lengths (1k-10k, 10k-20k, 40k-50k, 90k-100k, and 100k-300k bases) and assign 5Gb of reads into each of them (similar to coarse-grained workload in mm2-ax [15]). 2 Xeon is a trademark of Intel Corporation or its subsidiaries…”

Section: Experiments Setupmentioning

confidence: 99%

“…In particular, the chaining kernel in the seed-chain-align strategy of Minimap2 is an increasingly pronounced bottleneck for ultra-long reads. Time spent in chaining grows from 40% to 68% as average ONT read length grows from 2.25Kb to 127.11Kb [15]. This motivates us to accelerate Minimap2 for higher throughput on a GPU.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

mm2-gb: GPU Accelerated Minimap2 for Long Read DNA Mapping

Dong,

Liu,

Sadasivan

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Long-read DNA sequencing is becoming increasingly popular for genetic diagnostics. Minimap2 is the state-of-the-art long-read aligner. However, Minimap2's chaining step is slow on the CPU and takes 40-68\% of the time especially for long DNA reads. Prior works in accelerating Minimap2 either lose mapping accuracy, are closed source (and not updated) or deliver inconsistent speedups for longer reads. We introduce mm2-gb which accelerates the chaining step of Minimap2 on GPU without compromising mapping accuracy. In addition to intra- and inter-read parallelism exploited by prior works, mm2-gb exploits finer levels of parallelism by breaking down high latency large workloads into smaller independent segments that can be run in parallel and leverages several strategies for better workload balancing including split-kernels and prioritized scheduling of segments based on sorted size. We show that mm2-gb on an AMD Instinct(TM) MI210 GPU achieves 2.57-5.33x performance improvement on long nanopore reads (10kb-100kb), and 1.87x performance gain on super long reads (100kb-300kb) compared to SIMD accelerated mm2-fast. mm2-gb is open-sourced and available at https://github.com/Minimap2onGPU/minimap2.

show abstract

Section: Quadratic Complexity Of Chainingmentioning

confidence: 99%

Section: Algorithm 1 Segment Cutting In Range Selection Kernelmentioning

confidence: 99%

Section: Experiments Setupmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

mm2-gb: GPU Accelerated Minimap2 for Long Read DNA Mapping

Dong,

Liu,

Sadasivan

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Additionally, GPUs are becoming a more common choice for accelerated computing on sequencers– ONT sequencers GridION, PromethION, and MinION MK1C have GPUs built into them 7 . GPUs are also widely available in workplaces and on cloud platforms.…”

Section: Introductionmentioning

confidence: 99%

Accelerated Dynamic Time Warping on GPU for Selective Nanopore Sequencing

Sadasivan

Stiffler

Tirumala

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

The design and supply of RT-PCR primers for accurate virus testing is a complex process. The MinION is a revolutionary portable nanopore DNA sequencer that may be used to sequence the whole genome of a target virus in a biological sample. Human samples have more than 99% of non-target host DNA and Read Until is a protocol that enables the MinION to selectively eject reads in real-time. However, the MinION does not have any in-built compute power to select non-target reads. SquiggleFilter is a prior work that identified the accuracy and throughput challenges in performing Read Until using the state-of-the-art solution and proposed a hardware-accelerated subsequence Dynamic Time Warping (sDTW) based programmable filter on an ASIC. However, SquiggleFilter does not work for genomes larger than 100Kb. We optimize SquiggleFilter’s sDTW algorithm onto the more commonly available GPUs.DTWaxbetter uses tensor core pipes, 2X-SIMD FP16 computations and efficient data handling strategies using offline pre-processing, coalesced global memory loads, warp shuffles and shared memory buffering among other optimizations.DTWaxenables Read Until and yields 1.92X sequencing speedup and 3.64X compute speedup: costup over a sequencing workflow that does not use Read Until.

show abstract

Accelerating spliced alignment of long RNA sequencing reads using parallel maximal exact match retrieval

Wang,

Zhang

2024

Computers in Biology and Medicine

View full text Add to dashboard Cite

Accelerating Minimap2 for Accurate Long Read Alignment on GPUs

Cited by 15 publications

References 17 publications

mm2-gb: GPU Accelerated Minimap2 for Long Read DNA Mapping

mm2-gb: GPU Accelerated Minimap2 for Long Read DNA Mapping

Accelerated Dynamic Time Warping on GPU for Selective Nanopore Sequencing

Accelerating spliced alignment of long RNA sequencing reads using parallel maximal exact match retrieval

Contact Info

Product

Resources

About