Monday, May 11, 2020

COVID Sequencing: Is It About to Become Important for Public Health?

One of the great early victories in the COVID-19 battle was generating a complete sequence of the virus now known as SARS-CoV-2.   Soon, it was recognized that there were many emerging but clinically silent mutations.  For example, it was identified by March that most Italian cases may have come from Germany, and that most New York City cases may have come from Europe (see NYT lead story, April 8, here.)

May 10, the British newspaper The Guardian ran two articles about COVID-19 sequence mutations.   The first, by science editor Ian Sample, is an overview for the educated public on COVID mutations (here).  The second is an article by Sample on potential clinical implications of adaptive mutations of CSARS-CoV-2 (here).    He keys off what I think is one of the largest global sequence variants publications, by Phelan et al., in pre-release at - here.  They built a new phylogenetic tree with 3,958 viral SNPs from a total library of 5,349 whole genomes.  

For example, complex genetic trees can be constructed (these do not necessarily have any clinical correlate):

And mutations can be mapped to points on the COVID spike, as below (green points):

Large Companies Discussing Sequencing

Large players in the genomics industry are discussing sequencing.   Thermo Fisher announced an RUO setup for COVID sequencing on its new Genexus platform (here, here).    Illumina's investor call last week discussed the potential market for sequencing, including very high throughput sequencing for mass diagnostics (here, here).   Guardant's investor call also highlighted  a potential move into COVID sequencing for what has historically been an oncology company (at Genomeweb premium here, original transcript here).   This is just a highlight of the scientific and economic activity that is starting to pour into the U.S. and global COVID sequencing infrastructure.  

Public Health Value: Contact Tracing?

CMS currently pays $100 for COVID PCR (or any method) high throughput analysis of COVID for diagnostics (code U0004).   At scale, COVID sequencing might come into the same price range, and provide both diagnostic and sequence information.  Alternatively, if the same sample can be used twice, large-scale testing could be run for every 100 samples by PCR, with 1 in 10 or 1 in 20 samples that are positive for PCR being shunted into a secondary SEQ test.   (For example, CMS would pay $10,000 for 100 PCR samples for diagnostics under U0004.  If 10 were positive and one were sampled for SEQ, and seq cost $150, the total cost impact of the single extra $150 over the $10,000 total payment would be very tiny).   

Look at what could happen - it's highly favorable.  Public health authorities would have nearly-automated ongoing visibility in to the COVID sequence profiles in a community (assuming over an intervals tens of thousands of total COVID tests are being run, the 1 in 10 sampling will be an accurate survey.)  And this approach could also provide a quality check for whether contact tracing is actually pairing up individuals who truly were infected by the same viral strain.   This could lead to rapid quality improvement and selection of the best (and most efficient) practices in contact tracing, something that will be mission-critical to the US economy in Q2, Q3, Q4 of 2020.   Without spot checking for sequence, public health official will be flying entirely blind whether they are truly tracking and matching up real carriers and the corresponding infected persons accurately.

For an update from CDC, see the homepage for its SPHERES consortium for COVID sequencing,


The day after I wrote this, Genomeweb (Julia Karow) ran a deep dive article on the growing importance of COVID sequencing (vs PCR testing) - here.

Update 2:

A few days later, Genomeweb ran an article (Christie Rizk) on active use of COVID SEQ testing in a hospital, identifying unexpected patterns of virus in a study including patients, family members, and hospital staff.  Here.