Each time the SARS-CoV-2 virus copies its RNA, mistakes creep in. Most of the time, these errors are silent or even harmful to the virus. Occasionally, however, the changes allow COVID-19 to infect cells more readily or better evade the immune response. These mutations are the ones that form the basis of new variants of concern, such as Delta and Omicron.
Scientists like Eric Chow, a geneticist at the University of California, San Francisco (UCSF), are working hard to find new variants as they begin to emerge. It’s a job where sequencing is crucial. I sat down with Chow to learn more about his work and how he conducts high-throughput sequencing.
How did you first find yourself sequencing COVID-19?
My involvement with COVID-19 started early on in the pandemic when there was very little testing available throughout the US. In the San Francisco Bay Area, public health departments could run several dozen tests a day. That was it.
The UCSF and the Chan Zuckerberg Biohub came together to create an emergency COVID-19 testing facility. We set the lab up, validated the assays, and trained technicians – all in six days. At first we did a lot of testing for health departments, clinics, and groups across California.
By November 2020, we had stopped doing diagnostic testing and started COVID-19 surveillance. Though we were doing more COVID-19 sequencing than was possible in early 2020, we wanted to see if we could increase throughput for surveillance testing. And that’s when our lab really got to work.
What challenges did you face with surveillance testing?
When we started surveillance sequencing, there were still supply chain problems with consumables – especially pipette tips. Some of the existing COVID-19 sequencing protocols used 10-20 tips per sample. In the end, we used a Beckman Coulter Life Sciences liquid handler that uses acoustic energy to transfer liquid without using tips. And that allowed us to get our protocol down to two tips per sample, which helped us scale up our operations, while avoiding supply issues. As a side note, there’s no way we’d be able to do the work we do now in a manual way; I think the staff would quickly develop carpal tunnel syndrome if they had to do so much with multichannel pipettes!
What other strategies did you use to increase throughput?
We developed an amplicon-based approach to sequencing COVID-19. Existing protocols use overlapping PCR amplicons to copy virus cDNA that are then sequenced, and then go through traditional library prep to create a barcoded sequence library. Instead, we decided to adapt a technique developed by scientists at the University of Minnesota, which added small adapter sequences to the tails of every single primer. After we do the first multiplex PCR, we can immediately do a second PCR to add barcodes to the adapters, which results in a more streamlined approach.
One downside to this method is that the amplicons were fairly long at 400 bases, which meant that we were limited to certain machines for sequencing – and not everyone has access to those sequencers. To address this, we developed a shorter amplicon set so that instead of sequencing 400 base pairs, we were sequencing 275 – short enough to go on any Illumina sequencer.
The shorter amplicon length also meant that we could perform a sequencing run in just one day instead of two, which clearly increased our throughput. Now, we get RNA from positive COVID-19 samples from the California Department of Public Health, and as soon as we have a full run of 768 samples, we start the process.
Why are all these strategies important when doing surveillance?
A major change for us was thinking less about how we can increase our success rates of recovering a genome from a sample, and instead focusing on how we can maximize the outcomes that we get for a given amount of money or a given amount of time.
If your goal is to get 100 percent of the genomes of every single sample on a 96-well plate, then it’s going to be an expensive and laborious project. But if you have a method that allows 80 percent of the samples on two 384-well plates – and sequence all 768 at once for the same cost and time – you’re better off with that higher throughput and lower success rate. In the end, you’ve sequenced more genomes, which is what really matters.
What have you learned about infectious disease surveillance?
If there is going to be a new variant – one that is more transmissible, more pathogenic, or that can escape immunity from vaccines and previous infections – we need to know about it as soon as possible. And that means that we need to do a lot of sequencing. Scientists can then use this information to conduct experiments on how well antibodies neutralize the new variants and determine whether the variant is concerning.
Looking at data from variants such as Alpha, Delta, and Omicron, genomic surveillance detected these variants at very low levels – probably weeks before they started driving waves of infection. If scientists see a new variant rising rapidly, it might give us an indication that this is something to look out for, but we have to sequence a large number of cases in a relatively fast manner.
Getting these data into the hands of scientists and public health officials is critical. Genomic surveillance creates an early warning system and, the more sequencing you can do, the sooner you can detect these variants.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
