DNA analysis of early human remains from a Siberian cave has revealed the existence of a mystery human species.
A team of researchers speculates that this could have been Homo erectus, which lived in Europe and Asia a million years ago or more.
Meanwhile, the researchers report that they have also obtained the most complete DNA sequence ever from a Neanderthal.
Details of the work appear in Nature journal.
Finds at Denisova cave in Siberia have deepened our understanding of the human groups living in Eurasia before modern humans (Homo sapiens) arrived on the scene.
The Neanderthals were already well known, but DNA analysis of a finger bone and a tooth excavated at the cave revealed evidence of a human type living 40,000 years ago that was distinct both from Neanderthals and modern humans.
When this work was published in 2010, the team behind the discovery dubbed this human species the "Denisovans" after the Siberian site.
The Neanderthal toe bone was found in the same cave in 2010, though in a deeper layer of sediment that is thought to be about 10,000-20,000 years older. The cave also contains modern human artefacts, meaning that at least three groups of people occupied the cave at different times.
A high quality genome sequence was obtained from the small bone using techniques developed by Prof Svante Paabo and colleagues at the Max Planck Institute for Evolutionary Anthropology in Leipzig, Germany, and it reveals some interesting insights about both the Neanderthals and other human types.
For example, the researchers say, the Neanderthal woman was highly inbred and could have been the offspring of half-siblings who shared the same mother.
Other scenarios are possible though, including that her parents were an uncle and niece or aunt and nephew, a grandparent and grandchild, or double first-cousins (the offspring of two siblings who married siblings).
Comparisons of the genetic sequence of multiple human groups - Neanderthals, Denisovans and modern humans - yielded further insights into their evolutionary relationships.
The results show that Neanderthals and Denisovans were very closely related, and that their common ancestor split off from the ancestors of modern humans about 400,000 years ago. The genome data reveal that Neanderthals and Denisovans diverged about 300,000 years ago.
But it also threw up a surprise result: that the Denisovans interbred with a mysterious fourth group of early humans that were living in Eurasia at the time. Between 2.7 and 5.8% of the Denisovan genome comes from this enigmatic species.
This group split from the others more than a million years ago, and may represent the early human species known as Homo erectus, which fossils show was living in Europe and Asia a million or more years ago.
But Spanish researchers also recognise a species known as Homo antecessor, whose fossils show up about a million years ago at the Atapuerca site, near Burgos in Spain, and this may be another candidate.
Though Denisovans and Neanderthals eventually died out, they left behind bits of their genetic heritage because they occasionally interbred with modern humans. The research team estimates that between 1.5 and 2.1 percent of the genomes of modern non-Africans can be traced to Neanderthals.
Denisovans also left genetic traces in modern humans, though only in some Oceanic and Asian populations.
About 6% of the genomes of Aboriginal Australians, New Guineans and some Pacific Islanders can be traced to Denisovans, studies suggest.
The new analysis finds that the genomes of Han Chinese and other mainland Asian populations, as well as of Native Americans, contain about 0.2% Denisovan genes.
"The paper really shows that the history of humans and hominins during this period was very complicated," said Montgomery Slatkin, professor of integrative biology at the University of California, Berkeley.
"There was lots of interbreeding that we know about and probably other interbreeding we haven't yet discovered."
As part of the study, Prof Slatkin's colleague Fernando Racimo was able to identify at least 87 specific genes in modern humans that are significantly different from related genes in Neanderthals and Denisovans.
This, the researchers say, may hold clues to behavioural differences distinguishing us from early human populations that died out.
"There is no gene we can point to and say, 'this accounts for language or some other unique feature of modern humans'," Prof Slatkin explained.
"But from this list of genes, we will learn something about the changes that occurred on the human lineage, though those changes will probably be very subtle."
According to Prof Paabo, the list of genes "is a catalogue of genetic features that sets all modern humans apart from all other organisms, living or extinct".
He added: "I believe that in it hide some of the things that made the enormous expansion of human populations and human culture and technology in the last 100,000 years possible."