This story was originally published on May 6.
Researchers who last week published a proteomics study on an 80 million year old dinosaur are hoping they have silenced the critics who cast doubt on their earlier work on a Tyrannosaurus rex fossil.
Last week, a team of researchers published a paper in Science that discussed how they detected and sequenced peptides in a duck-billed hadrosaur. The study followed similar work by the same team, published in Science in 2007 and available here and here, in which they sequenced the proteins in a 68 million year old T. rex.
While the current work drew far less attention than the 2007 studies, the hadrosaur findings may have even more far-reaching scientific consequences, serving to validate the T. rex work, and further opening up the possibility that molecular data can be used in paleontology.
"The purpose was to show [we could detect and sequence proteins] on a new bone, to show that it's possible on a second fossil from a second location and that the first time was actually not a one-hit wonder," John Asara, a co-author on all three studies, director of the Beth Israel Deaconess Medical Center Mass Spectrometry Core, and an instructor at Harvard Medical School, told ProteoMonitor last week.
When he and his colleagues published the T. rex studies two years ago, they generated a swirl of buzz for their unprecedented work in decoding the proteins in such an ancient fossil, and adding along the way new evidence that modern-day birds descended from dinosaurs.
But after the initial excitement, criticism started to gather concerning the quality of the research done by Asara and his co-authors and their conclusions. In one commentary published January 2008 in Science, a team of 27 scientists evaluated Asara's work using authentication tests developed for ancient DNA, and found that although their work analyzing proteins from mastodon samples "pass[ed] these tests, [the] absence of amino acid composition data, lack of evidence for peptide deamidation, and association of alpha-1 type 1 collagen sequences with amphibians rather than birds suggest that T. rex does not," pass the test.
"There weren't very many peptides there, the peptide quality was a bit poor, and therefore, the interpretation in the phylogenetic sense was not very strong," Matthew Collins, one of the authors on the commentary, told ProteoMonitor this week. He is a professor of archaeology at the University of York and a protein mass spectrometrist.
"Does the age and environmental history suggest survival [of the collagen]? It seems unlikely," he said. "Does the data sense that it wasn't damaged? Well, it actually looked too good, almost. [The collagen] didn't look damaged at all. And did the results look right in terms of the phylogeny? It didn't seem quite right."
In another commentary, one particularly vocal critic, Pavel Pevzner, a professor of computer science at the University of California, San Diego, analogized the work done by the T. rex team to a monkey typing randomly on a keyboard: Given enough chances, the monkey will type out six-letter groups that are actual words, but that does not prove that the animal can actually spell, Pevzner noted.
In the same vein, Pevzner suggested, Asara and his colleagues' work in generating seven peptide sequences in T. rex did not prove that those sequences were actually from the dinosaur because at the time the T. rex researchers had not revealed all their generated spectra. Without knowing this, the false discovery rate could not be determined and "without addressing the statistical significance problem, [their results] are no more convincing than the first sensational report of dinosaur DNA published … more than a decade ago," Pevzner wrote.
[ pagebreak ]
Eventually, Asara did release all the data but that created more questions. After re-analyzing the data, Martin McIntosh, a full member at the Fred Hutchinson Cancer Research Center, found "a lot of collagen" as Asara had, but also a "high-quality hit to a hemoglobin peptide" that was an exact match to ostrich, McIntosh told ProteoMonitor. The majority of collagen peptides that were described in T. rex were also exact matches to ostrich, he said.
Because Asara's lab also does proteomics work with ostrich samples, that raised the possibility that the T. rex findings were contaminated and that the spectra attributed to the dinosaur fossil, including the collagen detected and sequenced by Asara, as well as the hemoglobin, may actually have come from ostrich samples.
Daedalus Falls Back to Earth
The criticism put Asara and his colleagues on the defensive, and although they continue to stand by their original findings, they also worked assiduously to preempt any perceived errors and omissions in their current study, Asara said.
"I think we've learned from the criticism and the fact that these are extraordinary claims, how you need to report this stuff and all the extra steps that you need to go through," he said.
Along with extra sediment and buffer controls, the researchers used two mass-spec data algorithms — Sequest and Mascot — to search fragmentation spectra against the reversed SwissProt protein database. In the 2007 work, only Sequest was used, "and we had no statistical evaluation. All we had were raw database scores and we had synthetic peptide comparisons," Asara said.
Asara and colleagues also performed extra validation this time around, using scoring statistics, decoy databases, manual inspection, and spectral comparisons to high-confidence spectra from synthetic peptides and peptides of identical sequence from extant organisms.
Unlike in the earlier studies, the team used an Orbitrap XL from Thermo Fisher Scientific to analyze spectra from the hadrosaur fossil, along with an ion trap mass spec, allowing Asara to determine more sequences more confidently, he said, because of the higher mass accuracy.
They were able to identify 149 amino acids in the current paper, compared to 89 for T. rex.
They also had William Lane, who runs the microchemistry and proteomics facility at Harvard, run tests on the bone extracts "that were showing the most intense spectra" to verify some of the peptides. In total, Lane confirmed the presence of three of the eight collagen fragments that Asara's team identified.
The consensus view is that Asara and his colleagues got it right with the hadrosaur study. Said the University of York's Collins this week, "I'm more convinced than I was after the first paper," that Asara and his colleagues found collagen in the fossils.
Importantly, there was evidence of protein deamidation this time around, and while the conventional belief is that collagen cannot survive 80 million years, "the more examples that come up of collagen surviving in materials of this age [and] the more times the team is able to get collagen out, the more it begins to suggest, in fact, our understanding of the way collagen preserves is wrong," he said.
[ pagebreak ]
While Collins would like to have seen an outside lab extract and analyze the whole sample — and not just some of the bone extracts as was done by Lane — he said, "I think there's pretty much no question now that the new mass spec data … really very much look like we've got collagen."
Pevzner added, "I think this study is a leap forward [from] the previous paper, which was computationally illiterate, essentially. … They've done now professional work."
Pevzner, however, remains unbending about the lack of clarity from Asara on the hemoglobin issue brought up by McIntosh.
Based on McIntosh's work — which was submitted to Science but was turned down — there is proof enough of the existence of hemoglobin in the T. rex sample, Pevzner said. And there can be only two explanations for this: The first is that the hemoglobin proteins detected by McIntosh are contaminants, "thus implying that the whole T. rex study is compromised and should be withdrawn," Pevzner said.
On the other hand, if the hemoglobin proteins, in fact, belong to T. rex this would represent a "breakthrough discovery in the area of T. rex proteins, implying that dinosaur bones represent a treasure trove of proteins not limited to collagens," he said, adding that he finds the researchers' silence on the hemoglobin issue puzzling.
"If it's coming from T. rex, they should be screaming about this … it's a great discovery. It means that they found [and] proved the existence of proteins in ancient bone that are not only collagen but hemoglobin and who knows what else," Pevzner said.
Asara said that his team found evidence of hemoglobin in the 2007 work but didn't address it at the time because it was only a single sequence and there was not enough data to write about it. In the hadrosaur study, Asara again found some evidence of hemoglobin, though only through antibodies and not sequences. Again, the researchers didn't feel they had enough data to include it as part of their findings, he said.
However, "now that we have evidence that we likely have hemoglobin … in the new bone, I think we have proven that at least we've gone in the direction to actually prove the stuff is endogenous," he said, adding that his team is currently trying to retrieve more hemoglobin proteins from the hadrosaur.
McIntosh said that though the current study doesn't directly answer questions about the existence of hemoglobin in T. rex, the fact that it now offers strong proof of the existence of collagen in the hadrosaur fossil suggests that contamination did not occur in the T. rex findings, and that the hemoglobin that he found is more likely to have been endogenous.
He added that because the hemoglobin peptide has never been published, it is not practical for Asara to address it in print.
"Based on the first experiments, I think the world felt that it was up to Asara and [his colleagues] to do more due diligence to rule out alternative explanations," McIntosh said. "I think with this new finding, it's becoming more and more and more of critics [who have to] provide evidence" that contamination took place and what was detected by Asara and his team wasn't collagen from the fossil.