Evolutionary dynamics of SARS-CoV-2


Trevor Bedford

Fred Hutchinson Cancer Center / Howard Hughes Medical Institute
24 Jun 2023
John J. Holland Lecture
Symposium on Understanding How Viruses Spread and Evolve
ASV Annual Meeting
Slides at: bedford.io/talks

Population immunity has driven a dramatic shift in death toll

But this immunity is now driving evolution of the virus

In which new variants emerge that escape from existing population immunity and spread rapidly

Novel variants sweep globally in months rather than years

Influenza H3N2

We really should be at something like Psi at this point

1. SARS-CoV-2 genome evolution

2. Variant frequency dynamics

3. Evolution driving epidemics

4. Forecasting

5. Continued evolution

SARS-CoV-2 genome evolution

Phylogeny of 2k viruses sampled globally in June and July 2022

Measure clade growth as a proxy for viral fitness

Across 2020 to 2023, clades with more S1 mutations grow faster

Rate of amino acid evolution and correlation between mutation and clade growth strongest in S1

Strength of adaptive evolution consistently high through time

Mutations at spike S1 propel escape from population immunity

These mutations are accumulating just as quickly post-Omicron

And are accruing much more rapidly than other endemic viruses

Variant frequency dynamics

Population genetic expectation of variant frequency under selection

$x' = \frac{x \, (1+s)}{x \, (1+s) + (1-x)}$ for frequency $x$ over one generation with selective advantage $s$

$x(t) = \frac{x_0 \, (1+s)^t}{x_0 \, (1+s)^t + (1-x_0)}$ for initial frequency $x_0$ over $t$ generations

Trajectories are linear once logit transformed via $\mathrm{log}(\frac{x}{1 - x})$

Variants show consistent frequency dynamics in logit space

Variants show consistent frequency dynamics in logit space

Multinomial logistic regression

Multinomial logistic regression across $n$ variants models the probability of a virus sampled at time $t$ belonging to variant $i$ as

$$\mathrm{Pr}(X = i) = x_i(t) = \frac{p_i \, \mathrm{exp}(f_i \, t)}{\sum_{1 \le j \le n} p_j \, \mathrm{exp}(f_j \, t) }$$

with $2n$ parameters consisting of $p_i$ the frequency of variant $i$ at initial timepoint and $f_i$ the growth rate or fitness of variant $i$.

Variant frequencies across countries from Feb 2022 to present

We find that recent variants like XBB.1.5 are ~300% fitter than original Omicron BA.1

Evolution driving epidemics

Many fewer reported cases in England post-Omicron

Data from UKHSA

ONS Infection Survey provides rare source of ground truth

Roughly 1 in 3 infections detected in 2021, while 1 in 40 in 2023

Data from ONS

Partitioning ONS incidence based on sequencing data shows variant-driven epidemics

~110% population attack rate from March 2022 to March 2023

~27k deaths in ~61M infections yields IFR of 0.04%

Data from UKHSA and ONS


Assessing MLR models for short-term frequency forecasting

MLR models generate accurate short-term forecasts

Now have clade and lineage forecasts continuously updated

Multinomial logistic regression should work well for SARS-CoV-2 prediction, except new variants have been emerging fast enough that the prediction horizon is really quite short

Could we predict the spread of new mutations using DMS data?

Escape from antibodies that potently neutralize BA.2

Can calculate escape of arbitrary RBD against antibodies known to neutralize BA.2

Strong correlation between DMS immune escape and lineage-level MLR growth advantage

Continued evolution

How likely are further "Omicron-like" events?

With 1 observation in 3.6 years of virus evolution, it's currently unclear how common Omicron-like events will be

This rate distribution gives a naive prediction of ~23% probability of an Omicron-like event occurring in the next 12 months

1 in 4 seems high, but "cryptic lineages" are commonly being found in wastewater

Even without further "saltational" events, ongoing evolution is rapid and so far shows little evidence of slowing

This pace of evolution contributes to large burden of disease at endemicity

  • Assume yearly attack rate of ~50% (even without the potential for Omicron-like emergence this may be an underestimate)
  • Assume IFR of 0.05% (similar to influenza, compare to early pandemic IFR of 0.5%)
  • This would give somewhere around 80K yearly deaths in the US at endemicity
  • For comparison, this year the US has seen 40k COVID-19 deaths from Jan 1 to Jun 10, which would extrapolate to 90k deaths in 2023

Research priorities to deal with this rapid evolution

  1. Continued genomic surveillance and evolutionary forecasting
  2. Development of new antivirals
  3. Vaccination strategies to combat antigenic evolution and OAS
    • Universal vaccines?
    • Cocktail vaccines?
    • Masking epitope sites to reduce imprinted response?


John J. Holland Lectureship

SARS-CoV-2 genomic epi: Data producers from all over the world, GISAID and the Nextstrain team

Bedford Lab: John Huddleston, James Hadfield, Katie Kistler, Thomas Sibley, Jover Lee, Cassia Wagner, Miguel Paredes, Nicola Müller, Marlin Figgins, Victor Lin, Jennifer Chang, Allison Li, Eslam Abousamra, Donna Modrell, Nashwa Ahmed, Cécile Tran Kiem