The fungus killing frogs around the world comes in several forms, and has almost certainly been distributed by trade in amphibians, research shows.
Scientists led from Imperial College London found three distinct lineages of the chytrid fungus in various nations.
The most widespread and lethal form was probably created by a crossing of two prior forms, they report in Proceedings of the National Academy of Sciences.
Chytrid is now found on every continent and has wiped out a number of species.
Identified just over a decade ago, it kills amphibians by blocking the transfer of vital substances through their skins, eventually causing cardiac arrest.
Its origins are believed to lie in southern Africa.
"Before this study, no-one knew there were any different lineages," said Rhys Farrer, the project leader from Imperial.
"This work comes from using the new whole-genome sequencing technique, combining data from all over the world.
"And it's obviously important, as chytrid is one one of the most devastating wildlife diseases with the largest host range of any, and responsible for dozens of species extinctions and many more extirpations of local populations."
Cape of good hope
The team took samples from amphibians in 20 sites spanning Europe, North America, Central America, the Caribbean and South Africa.
The majority carried Batrachochytrium dendrobatidis (Bd) fungus of the type that has a truly global spread, which they dubbed BdGPL.
But their Swiss sample showed a different form, or clade, named BdCH, while a third clade (named BdCAPE) turned up in the Cape Province of South Africa and the Mediterranean island of Mallorca.
The Mallorcan chytrid was almost certainly carried from South Africa, probably via the trade in amphibians for zoos or private collections.
The Swiss form probably came via a similar route, researchers believe.
Laboratory tests showed that BdCAPE was substantially less damaging to amphibians than BdGPL. (The Swiss form was identified too late in the project to be tested in this way.)
The genetic differences that make BdGPL more lethal have not been identified. But the team believes it became so deadly through a chance encounter between two or more prior strains.
"We think we are seeing unique evidence of recombination within BdGPL - we can't say for sure if it's a hybridisation event but it's the most likely explanation," said Mr Farrer.
"From the dating work we've done it's safe to say that it arose in the 20th Century, and that's in the realm of time for the trade in amphibians."
Although the transport of exotic amphibians for pets is a prime suspect, another theory holds that the lethal BdGPL chytrid spread through the importation of frogs from Africa to North America and Europe for use in pregnancy testing.
However, yet another form of the fungus was recently discovered in Japan, its relationship to African-derived lineages uncertain.
The latest research marks a new staging post on a fast and fascinating voyage of scientific discovery.
Whether it can help combat the disease is another matter.
Analysing the genomes of the various strains may show scientists what makes some virulent and others relatively benign.
The Imperial team believes it is also worth investigating whether the less virulent forms can be used to give amphibians a degree of resistance, in the same way that some vaccines do through using attenuated forms of disease-causing microbes.
Follow Richard on Twitter