Genomics Coming of Age: Accelerating A Cure – Part I of 2

By Julie DiBene

Director, Marketing Communications

MoSys, Inc.

COVID-19. Seems the more we learn, the more we learn that we don’t know a whole lot of about this global killer. The medical community is scrambling just to keep up with new cases as the world waits anxiously for both a viable treatment and a vaccine. It is somewhat ironic that even people who have never given a moment’s thought to the discipline of genomics are now reading about possible breakthroughs and cures on a daily basis. No doubt, genomics has come of age.

Even before COVID-19 changed the course of society and history, the need for computation related to very large data sets in modern biology had made significant advancements in both volume and complexity. For instance, high-throughput gene and DNA sequencing has already been used in a wide range of topics in biology and medicine. These require ever finer analyses of data such as personal genomics profiling for medical applications, including diagnosis and disease monitoring. Basically, the medical community has had a bit of a head start when it comes to sorting through the complexities of a virus like COVID-19.

As an example, in genomics research/diagnostics, stored DNA sequence data are retrieved and compared in order to check for the existence of a disease-associated gene traceable back from a given individual. In addition, sequencing of mRNA and the epigenome are mapped back to a reference genome, yielding millions of data points which need to be further compared with a reference genome. The genome sequences can be represented as graphs that represent various permutations. The analysis of these graphs has turned the memory access needed to read and analyze data in comparison to the reference genome into a significant bottleneck. While DRAM-based memory can store a lot of data, the random read bank activation rate becomes the limiting factor. Other memory classes can be considered such as SRAM, but the available storage density becomes another issue to be considered.

More recently in genomics prediction, classification and correlation are being handled by using an algorithm such as a Random Forest of Trees to help recognize patterns in the genome sequence graphs. When using Random Forest of Trees to address these challenges, systems are limited by the memory random access rate of the associated hardware that is running the algorithm. With the amount of data constantly increasing there is an ongoing need for increased processing power along with the possible use of a specialized hardware solution with efficient access to memory and localized processing that seamlessly integrates and efficiently accelerate these algorithms.

As one may expect, both Random Forest of Trees and DNA sequencing may prove vital in combatting COVID-19. Specifically, medical experts are trying to figure out why some people get a cough and are fine within a few days and others die, why still others carry the virus and never get sick and then others end up on ventilators in the ICU. Experts think the answers lie deep within patient’s DNA and how each person’s immune response manifests. Researchers have already zeroed in on human leukocyte antigen (HLA) genes, which contain instructions to build proteins that bind to the pathogen. These are the red flags that warn our immune cells. In turn, human immune cells, once they learn to identify the pathogen, kickstart the process of building antibodies to target and destroy the invasive germ. 

Thus far, researchers know that within each person, the HLA genes hold the code for at least three different classes of proteins. So, depending on which HLAs you already have,  your body may be able to effectively fight off certain infections — including SARS-CoV-2, the virus that causes COVID-19. Or not.

In Part 2 of this blog, we will delve deeper into how MoSys Graph Memory Engine (GME) technology can accelerate genomics and ultimately, a cure.

Additional Resources:

Share on social media