Tracking virus mutations

May 19, 2020 9:23 AM

Microbiologists: Given the host's role in virus reproduction could we theoretically track (with some statistical accuracy) individuals that a particular virus variant has passed through? For example, if 20 infected people visited the old folks' home, could we pick out the one that passed the virus to the residents?
posted by Tell Me No Lies to Science & Nature (6 answers total)

Nextstrain.org has been doing genomic sampling and tracking at a larger scale (global, North America) but based on their situation reports, your hypothetical tracking scenario would be very difficult unless there is a very distinct and timely mutation that happened. (Not a scientist, just follow them on Twitter)
posted by spamandkimchi at 10:08 AM on May 19, 2020

I think they've done this a few times. Ex: genetic lineage tracing to identify where clusters come from. They can tell if a particular cluster came from someone who was in Italy, or from someone who was in Washington. If the 20 visitors were all carrying the same strain, it would be hard to differentiate (like trying to have a lineup where all the suspects are siblings or twins). But if they were from different communities and had more steps in the transmission chain between them, it would become easier.

This requires that you have actual traceability, though - sequence data on the virus from the folks who were possible transmitters (or from their chain of transmission or cluster, who would have related viruses), and from the nursing home.
posted by Lady Li at 12:47 PM on May 19, 2020

Response by poster: I’m not sure exactly what makes for a new strain. My understanding is that there are minor copying errors every time reproduction takes place. I can see where over many generations things would get very muddled, but if it was just one or two copy cycles could you make a much more specific comparison?
posted by Tell Me No Lies at 2:05 PM on May 19, 2020

I reccomend emailing the folks of This Week in Virology. The site is microbe.tv
posted by kathrynm at 3:38 PM on May 19, 2020

Best answer: If it were 1-2 copy cycles you definitely couldn't. (Although you'd get far more copying than that from infection to infection.) Coronaviruses are weird RNA viruses in that they have an enzyme (an exonuclease) that proofreads their RNA as it's copied. This is a good thing for COVID-19 vaccine development, as it means SARS-CoV2 is mutating at a much slower rate than say, seasonal flu.

Unfortunately most of the estimates I've seen of rate of mutation are annualized, which to my non-epidemiologist, RNA biologist brain doesn't necessarily make the most sense for forward-extrapolating in a global pandemic where the number of cases (and thus, replication cycles) are hockey sticking upwards. I should really look for NextStrain's estimate (I'm sure it's on Trevor Bedford's Twitter), but working off the idea that the virus accumulates 1.3 x 10^-3 mutations per nucleotide per year (PDF cite) with a ~30,000 nucleotide long genome, you'd expect 39 new mutations annually. Basically, the answer is it's mutating much slower than one mutation per case.
posted by deludingmyself at 3:39 PM on May 19, 2020 [1 favorite]

Best answer: What you're describing is an actual tool we use in molecular epidemiology (which might be a good keyword/phrase for you to keep searching for more information). As deludingmyself notes above, though, the relative bang for the buck in very short time frames for infectious diseases is quite limited, and, in instances where point mutations might be what we're talking about, all sorts of problems with trying to trace serial passage through infected people can come up (including mutations reverting, re-reverting, and on and on). It can be surprisingly difficult to do with individuals (A gave it to B gave it to C) what it's surprisingly easy to do with big groups and populations (it started in region A, moved to region B, then to region C).
posted by late afternoon dreaming hotel at 3:48 PM on May 19, 2020 [1 favorite]