GGG Ping.
Makes you kind of wonder about the accuracy of any gene sequencing....
Here’s this, from Ars Technica:
By John Timmer | Published: October 15, 2007 - 10:33AM CT
About a year ago, we reported on the latest results in the sequencing of Neanderthal DNA, which had netted a million bases of the genome of our extinct relative. Those were just the first results of a sequencing effort that was planned to continue; there is probably about another 18 months worth of data that have accumulated since the writing of those papers had started. Unfortunately, an Open Access publication in PLoS Genetics suggests that at least some of the original results are artifacts, the result of human contamination.
The new analysis was triggered by differences in the methods and results in the two papers that were published on the topic. A paper in Science described cloning Neanderthal DNA in bacteria, and then sequencing it. The one in Nature sequenced amplified DNA directly. Although there was substantial agreement between the two, the Nature paper suggested that modern humans and Neanderthals shared a common ancestor more recently, and may have interbred after their separation.
The authors of the PLoS paper obtained the original sequence data from both papers, and performed a new analysis of it. They discovered that the data in the Nature paper contains sequence differences that appear to have arisen recently within the human lineage, which suggested something was wrong. Performing an estimation of the human-Neanderthal split date using the Nature data produced a value of 35,000 years, which is completely incompatible with the fossil record. Finally, using a date of 350,000 years for the split (obtained using the data from the Science paper), they found that the Nature data indicated extremely high levels of interbreeding between humans and Neanderthals; the Science data continued to suggest there was none.
All of these results point to one conclusion: the Neanderthal sequence in the Nature paper looks far more like that of modern humans than any other data would suggest is possible. Of course, there’s a simple and obvious explanation for that discrepancy: the sequence is from modern humans.
The new paper explores this via the following reasoning: ancient DNA is more likely to be damaged and fragmented, and so contamination is more likely to appear in longer, less damaged fragments. They divided the Nature sequence data according to the length into short, medium, and long pools. The short fragments give an age estimate for the Neanderthal-human split that’s essentially identical to the one obtained with the Science data. But, as the fragments get longer, the age shrinks. When fragments greater than 100 bases long are examined, they give an age estimate for the split that is younger than some splits within modern human populations. Thus, the longer fragments are very likely to be contamination from modern humans.
The authors recognize that more work needs to be done to sort out some remaining discrepancies, but the new analysis strongly suggests that a large portion of the original data was the result of contamination. We know less about the Neanderthals than we thought we did. Still, the analysis suggests that there is real Neanderthal sequence among the contaminants, and suggests a fairly simple analysis may help us extract it. It’s a great example of how science can self-correct.