MemVerge and Analytical Biosciences Accelerate Cancer and Covid-19 Research With Big Memory

MemVerge, Inc. and Analytical BioSciences Ltd. (ABio), provider in single cell genomics, announced they have accelerated the time-to-results for single-cell RNA sequencing (scRNA-seq) analysis by using Big Memory technology.

A bleeding edge solution composed of Intel Corp.’s Optane Persistent Memory (PMEM) and MemVerge Memory Machine software produced up to 800x faster scRNA-seq load times and 25x faster execution for some computational stages, compared to a traditional DRAM-based solution.

Scientists rely on genomic sequencing to identify new SARS-CoV-2 variants and understand how they will impact health. On February 17, The White House announced The Centers for Disease Control and Prevention (CDC) will invest nearly $200 million to identify, track and mitigate emerging strains of SARS-COV-2 through genome sequencing.

“Single-cell RNA sequencing is one of the key fundamental research methods that drives advances in Cancer and Covid-19 research. We are proud to be a world’s leading provider of the most advanced single-cell analytical tools,” said Chris Kang, head, bioinformatics operations, Analytical Biosciences. “The Big Memory platform that MemVerge and Intel developed accelerates our workflows and helps us generate results much faster, which will lead to more efficient ways to gain greater insights and knowledge in diseases mechanisms and improve healthcare.“

Single-cell RNA sequencing analysis is a process containing multiple large scale data processing steps. The long-running process is compute-intensive and imposes significant demands on memory resources. Computation uses very large matrices that need to fit in memory and intermediate results must be saved and reloaded for other stages, which introduces storage and recovery bottlenecks. Furthermore, the process has many stages that must be repeated for parameter tuning.

Memory Machine software from MemVerge addresses these memory resource issues. Its ZeroIO memory snapshots eliminate the storage IO bottleneck to reduce overall time to execute the entire scRNA-seq analysis. Other bioinformatics analyses that use large matrices derived from next-gen sequencing techniques can also be accelerated with Big Memory.

“Until now, memory infrastructure did not offer a viable alternative to storage for genomic sequencing,” said Charles Fan, CEO, MemVerge. “Big Memory offers the same high-performance as DRAM at a dramatically lower cost, and with the persistence and agility needed for complex data pipelines.“

scRNA-seq testing results
As a baseline, the complete analysis using R-based tools was completed using DRAM only. The analysis was repeated using Memory Machine and the combined DRAM plus PMEM memory pool. Each analysis stage included data loading and saving processes to mimic the real working scenarios. The testing showed that execution time was up to 25x faster with Memory Machine, saving a total combined execution time of 60% across all stages.

Click to enlarge

Testing also showed that data load times were up to 800x faster with Memory Machine
Click to enlarge

The bigger memory capacity allows more tasks to run in parallel and easy rollback to run what-if scenarios. These benefits combined with faster execution times resulted in an overall 10x increase in project throughput per server.

Resource:
How Big Memory is speeding up genomics research: Big Memory Accelerates Single-Cell RNA Sequencing