As a world-leading authority on virus evolution, Eddie Holmes wasn't surprised when COVID-19 first appeared.
Like many people who study the emergence of infectious diseases, Professor Holmes believes SARS-CoV-2 was an accident waiting to happen.
But in the years since he became one of the first people in the world to publicly share the SARS-CoV-2 genome, much of the way the virus has evolved has taken him by surprise.
"The level to which it's become more infectious is something that I would never have predicted," said Professor Holmes, a virologist at the University of Sydney.
This week marks three years since the World Health Organization declared COVID-19 a global public health emergency.
In that time, the virus has spawned five variants of concern, each more transmissible or potent than the last.
Omicron, which emerged in late 2021, has since splintered into hundreds of subvariants — amounting to what's been described as a "variant soup" — and will likely keep splintering into more.
"The virus has still got the capacity to evolve and it will evolve … there's no evolutionary dead-end," Professor Holmes said.
"We'll definitely see more immune escape variants … but when, where and what they'll be is very hard to say."
Splintering virus reflects 'immunologically complex' population
While Omicron continues to dominate globally, the current spread of SARS-CoV-2 looks different depending on where you are.
"We're seeing this complicated global picture of different places that have different immune profiles given their history of infection and vaccination," Professor Holmes said.
"That's very unlike what was happening in 2020 and 2021, when basically no-one was infected, so the virus kind of had free rein."
Earlier in the pandemic, the same variants tended to appear everywhere because everyone was in the same position: immunologically "naive".
These days, against a backdrop of divergent immune profiles and waning immunity, hundreds of Omicron subvariants have emerged as a result of geographically unique evolutionary pressures.
"That's why we're seeing these variant 'soup' events because they all come from different populations," Professor Holmes said.
"It makes predictions [about the virus] even harder because the global population is so immunologically complex now."
Australia has its own "unique" Omicron soup, according to virologist Stuart Turville, which is currently dominated by two subvariants: BR.2.1 and XBF.
In recent months, the emergence of new, more immune-evasive subvariants has triggered a huge wave of infections and reinfections.
Dr Turville, from the Kirby Institute, said although the subvariants in circulation were genetically "very diverse", there was no evidence to suggest any were causing more serious disease.
"In the context of how they enter our bodies and how they grow in our bodies, all of the Omicrons are pretty similar," said Dr Turville, whose team studies the distribution of SARS-CoV-2 in Australia.
But health experts have warned that it's still worth trying to avoid any new infection to reduce the risk of long COVID, and for vulnerable individuals, the possibility of severe disease.
The risk of importing new variants
Globally, COVID infections and deaths have been rising since early December, including in the United States, where the highly contagious XBB.1.5 subvariant — unofficially dubbed "Kraken" — has quickly outpaced the rest.
"A lot of people are talking about XBB.1.5 because it's got a fitness gain, in that it can bind to its primary receptor — ACE2 — to a greater affinity, and alongside that, it's very adept at navigating antibody responses," Dr Turville said.
But the strength of a subvariant in one country doesn't necessarily mean it will have the same advantage in another.
So far, only a small number of XBB.1.5 cases have been detected in Australia.
Epidemiologist Jodie McVernon said Australia appeared to be moving away from "big boom and bust outbreaks" to "a more endemic state", but warned that monitoring new subvariants remained key.
"We need to really closely monitor whether a strain that is dominant in one place will have an advantage in another," said Professor McVernon, who is director of epidemiology at the Doherty Institute.
"It's the sort of thing researchers are looking out for … and scanning for variants of concern to see if they have obvious immune escape properties."
In China, the unprecedented surge of COVID-19 cases since the country abandoned its zero-COVID approach has fuelled concerns that unrestrained, rapid transmission could give rise to a new variant of concern.
Generally, the more a virus spreads, the more opportunities it has to evolve.
But Professor Holmes said the fact the Chinese population was so susceptible to COVID-19 — due to lack of prior immunity — meant the virus wasn't under pressure to mutate.
"In that kind of naive population, you wouldn't see a strong selection pressure because the virus doesn't need to be selective, as there's no immunity in the population to stop it," he said.
"[The virus] wasn't going to select for a new variant … just the fastest strain."
The next 12 months
When asked what might happen over the next 12 months, Dr Turville said he didn't like to "crystal gaze" about what variants may or may not do. But if he were to make a guess, he said the virus was likely to follow one of three trajectories.
The first — and most optimistic — scenario would see the vast majority of Australians get up to date with their booster vaccines, including the newly available bivalent shots, which would help to "put the virus under pressure".
"Often, what [the virus] does to get around the antibody response is that it may need to drop something that gives it a fitness gain," he said.
In other words, the virus might need to make some kind of biological trade-off because it can't always do everything simultaneously well.
"We might have a situation where we can slowly but surely corner the virus into being less fit … and in that context, we'll hopefully see the reproductive rate go down, so [the virus] causes smaller waves, and puts less pressure on hospitals."
The second scenario, which he believes is most likely, looks not unlike our current COVID-19 situation: lagging booster rates and "not a lot of political discussion" means we'll face "an era of complacency", he said.
"When a new [variant] enters the community, it isn't under so much pressure and it has the ability to spread … and that might drive waves bigger than what we would expect.
"Unfortunately, it will create situations where people that are vulnerable may become infected and that may lead to complications."
In the third, "worst case" scenario, the virus might evolve into something that we can't predict.
"We might find the virus seismically shifts … a variant pops up that is incredibly immune evasive and can drive a very large wave," Dr Turville said.
"We may also have a situation where we might see increased disease severity … But I'd say from what we're seeing in the variants that we're doing surveillance on, that last scenario has the lowest probability."
Professor Holmes agreed that the virus was unlikely to become more virulent, but said he wouldn't rule anything out.
"What we've seen with Omicron is a virus that is less able to infect the deep lungs, because it's got more specialised in the cells of the upper respiratory tract and that enhances transmission.
"If it continues that process, you'd expect it not to evolve back to being a more virulent thing."
Will COVID become seasonal?
While until this point, COVID-19 waves have been mostly "globally synchronised", Professor Holmes said increased immunity means we might start to see more seasonal patterns emerge.
Approximately 70 per cent of the world's population has now received at least one dose of a COVID-19 vaccine.
"The more immunity you have in the population, the harder it is for the virus to spread, therefore optimal climatic conditions for its spread become more important."
In Australia, Professor McVernon said she expected to see some degree of ongoing transmission with "some seasonal outbreaks from time to time, a bit like the flu".
But Dr Turville said he wasn't convinced COVID-19 could be thought of as a seasonal virus yet.
"There's a lot of things that we're still learning in real time about the virus and the consequences of the virus infecting different populations," he said.
"Although we'd like to draw analogies to things like the flu and talk about it in predictable and seasonal terms, I don't think we're quite there yet."