Despite many blaming bats for COVID-19, a new study has found they're perhaps being unfairly targeted - a bat coronavirus simply couldn't make the jump.
SARS-CoV-2 - the version which spreads in humans and causes COVID-19 - is 1000 times better at binding itself to our cells than RaTG13, its closest relative found in bats, scientists in Australia and the UK say.
"SARS-CoV-2 binds very well, while the bat virus does not," said Francesca Di Giallonardo, a postdoctoral researcher at the University of New South Wales' Kirby Institute.
Rather than simply evolving to make the jump from bat to human, scientists say SARS-CoV-2 appears to be the result of recombination - when two different viruses infect the same cell and produce offspring with RNA or DNA from both. This can happen naturally or be done in a lab, though there's no suggestion in the new study that's what happened here.
Nikolai Petrovsky of Flinders University said the bit of the virus which hooks into our cells - the spike protein - looks like that found on pangolin coronaviruses. Both bats and pangolins were sold in the wet market in Wuha, China, where the virus was first reported.
"This pangolin CoV-like spike protein provided a presumed bat virus such as one mimicking RaTG13 with the ability to efficiently infect human cells, making it a highly efficient human pathogen."
Despite sharing about 96 percent of their DNA, previous research by Dr Petrovsky has found it would take 50 years of evolution for RaTG13 to turn into SARS-CoV-2 (and a 4 percent difference in DNA is huge - about the same as the difference between a human and a chimpanzee - but viruses mutate much more quickly thanks to their high rate of reproduction).
But if the pangolin virus had a spike protein ideal for infecting humans, why aren't they considered the likely source? Because the scientists found a part of the spike protein, the furin cleavage site, didn't look like those found in pangolins - or bats, for that matter.
"Without its pangolin CoV-like spike protein with its non-pangolin-like furin cleavage site, the SARS-CoV-2 virus could not have become a virulent human pathogen," said Dr Petrovsky.
"This makes it all the more critical to determine where the SARS-CoV-2 virus initially came from and how it was created, so that future pandemics with similar viruses can hopefully be prevented."
Another complication is that RaTG13 - the closest known relative of SARS-CoV-2 - might not even exist.
"It is important to appreciate that RaTG13 itself is a conceptual rather than proven real virus, as a virus corresponding to RaTG13 has never been isolated and cloned," said Dr Petrovsky.
"Instead, the sequence of RaTG13 was derived from short reads of amplified sequences isolated from pooled bat specimens, and thereby may itself be an artefact rather than a real virus, made up of sequences from different but related coronaviruses."
The Wuhan Institute of Virology had an RNA sequence of RaTG13 for research, but has denied reports it had a copy of the actual virus.
The scientists behind the latest study, published in journal Nature Structural & Molecular Biology, hope their findings help with the development of a vaccine. The death toll worldwide to date is 571,000.