Published online 27 February 2009 | Nature | doi:10.1038/news.2009.119


Origin of 'T. rex' protein questioned

Dinosaur samples may have contained traces of ostrich.

T. Rex skeletonWere samples of protein from a fossilized T. Rex contaminated?Wikimedia / David Monniaux

More doubt has clouded claims that dinosaur protein has been sequenced. Now a long-time critic has called for an independent review of the 2007 studies of ancient protein from a fossilized Tyrannosaurus rex after fresh analysis revealed traces of ostrich haemoglobin in the original samples.

In the contentious papers, researchers identified seven fragments from a protein called collagen1, found in connective tissue, and said their sequences most closely matched the chicken version of the protein2. The samples came from the fossilized femur of a T. Rex. As well as further strengthening the evidence for the link between dinosaurs and birds, the findings would make the protein the oldest ever to be sequenced — by around 68 million years.

The work, published in Science, garnered headlines worldwide and met with considerable scepticism at the time.

Now Martin McIntosh, a mass spectrometrist at the University of Washington in Seattle, says he has identified ostrich haemoglobin protein in a cache of 48,000 protein spectra released last September by John Asara, whose lab at Beth Israel Deaconess Medical Center in Boston analysed the original samples. McIntosh suspects the samples were somehow contaminated with proteins from modern species.

But Asara and colleagues are standing by their results and reject claims that their protein was tainted.

Dinosaur debate

McIntosh's findings came to light on 23 February at the sixth annual meeting of USHUPO — part of the international Human Proteome Organisation — in San Diego, California.

In the course of a lecture, Pavel Pevzner, a computational biologist at the University of California, San Diego, cited McIntosh's work and called for an independent review of the original results. He said his analysis provides statistical support for two of seven collagen proteins the Science authors said were from T. rex. But now the contamination issue has raised new concerns about the validity of the earlier findings.

It was not the first time that Pevzner had criticized the findings. Before Asara made the protein spectra available in September, Pevzner had publicly called for their release, arguing the statistical accuracy of the initial results couldn't be verified without them.

Pevzner and his colleagues also penned a critical technical comment on the statistical questions in Science last August3. In an earlier technical comment in January last year, 27 authors had written that they couldn't verify collagen proteins in the samples4, and a separate study last spring asserted the proteins are from slime mould5.

McIntosh also wrote a technical comment for Science, suggesting contamination on the basis the haemoglobin spectra — but it was rejected. Peer reviewers said that one fragment of ostrich haemoglobin was not enough to suggest contamination had occurred. McIntosh accepts that his paper doesn't prove contamination, but still believes the samples may be tainted.

Close to an ostrich

Asara is sticking by the findings, though his team reported in a follow-up paper in 2007 that there was no statistical support for one of the fragments6. He told Nature News it is "statistically unlikely" that only one ostrich peptide would show up if there was contamination, especially since sediment and reagents from his experiments had no trace of it. He claims that work he hopes to publish soon also rules against contamination.

Last May, he and Mary Schweitzer, the palaeontologist at North Carolina State University in Raleigh who provided him with the samples from the femur, published another Science article showing that further comparison of the ancient proteins to those of existing species — like ostrich and crocodile — confirmed their work7.


Meanwhile, the controversy rumbles on. When Pevzner polled his audience at the USHUPO meeting, none of the several dozen researchers present indicated support for the T. rex collagen results. Asara was in the room, but did not respond.

Marshall Bern, a computational biologist at the Palo Alto Research Center in California who was not at the meeting, said he leaned in support of Asara's collagen analysis. "Asara has met Pevzner's test," says Bern. "All this scrutiny makes the two proteins [that Pevzner verified] look good. And I'm not ready to discount the other five. Contamination is now the issue."

The two lead authors of the controversial papers say they may accept a review of the work — depending on its form.

"We agree independent verification of our findings is critical to their acceptance," Asara told Nature News, adding collaborations are already under way.

Schweitzer agrees. "If a review looked at everything, I would support it," she says. 

  • References

    1. Schweitzer M. H., et al. Science, 316, 277–280 (2007)
    2. Asara, J. M. et al. Science, 316, 280–285 (2007)
    3. Pevzner, P. A., Kim, S., Ng J. Science, doi:10.1126/science.1155006 (2008)
    4. Buckley, M. et al. Science 319, 33 (2008)
    5. Kaye, T. G., Gaugler, G., Sawlowicz, Z. PLoS ONE 3, e2808 (2008)
    6. Asara, J. M. et al. Science, 317, 1324–1325 (2007)
    7. Organ, C. L. et al. Science 320, 499 (2008)
Commenting is now closed.