Here's the key sentence in the article: "'We're looking at a very interesting population scenario'—one that does not jibe entirely with what we thought we knew about how waves modern human populations migrated into and through Asia and out to Oceania's islands."
Svante Pääbo was the pioneer of doing DNA sequencing on samples of old human tissues (he started with Egyptian mummies when he was an Egyptologist). I heard a lecture by Pääbo at the 2008 Nobel Conference
and heard there from some of the other researchers who use DNA analysis to trace human origins.
One of the problems in the discipline so far, as you can well figure out for yourself by thinking about it, is that the sample sizes for defining what genome pattern is "modern human" are still remarkably small, especially when we get to full genome sequences of individuals with well determined geographic origin, and smaller still for Neanderthal sequences and smallest of all for Denisovan sequences. So much work still needs to be done to figure out what the usual range of free variation of genes that are not under strong selection pressure is in the hominin genome line.
Figuring out what genes actually do at the individual level in a genome as complicated as the human genome has turned out to be much harder than was first supposed (as mentioned in the submitted article), so while tracing lineages with gene markers that are not under strong selection pressure helps define geographical population groups with common founder populations, knowing even an individual's complete genome does little to show the traits of the individual. (The first human genome sequenced didn't even reveal that the sequenced individual has blue eyes, for example.)
Turkheimer, E. (2012). Genome wide association studies of behavior are social science. In K. S. Plaisance & T.A.C. Reydon (Eds.) Philosophy of Behavioral Biology (pp. 43-64). New York, NY: Springer.
"If the history of empirical psychology has taught researchers anything, it is that correlations between causally distant variables cannot be counted on to lead to coherent etiological models."
"Most Genetic Associations with General Intelligence Probably False Positives"
"Presuming this is correct and, as Visscher and Keller appear to have implied, applicable to most behavioural traits, this indicates that literally thousands of genes are likely involved in each trait, with no single polymorphism having substantial effect, which is the quasi-in?nite model to which Munafo and Flint refer. Whatever would we as psychologists do with such information?"
AFTER EDIT: The interesting question raised by the top-level comment that was posted before this comment has received some good replies. The question, and the issue of defining a "species," reminds me of what Richard Dawkins points out about common descent. If you lined up in a line by order of descent with ALL of your ancestors, including the ancestors no longer living, as you looked at each individual in the line you could say "He is the son of the father standing next to him," going back as far as you want, even hundreds of millions of years, and yet at some point in the deep past the ancestors would not look anything at all like "humans," or indeed even like mammals or like tetrapods. At the individual level, OF COURSE you are part of the same species as your parents (by definition of "species"). Life on Earth today has a common ancestor, or at the very least a common set of one-celled ancestors, by general agreement of biologists, so all the species that have differentiated from one another over time can be traced to common ancestors of multiple species. You are undoubtedly related to your parents, who belong to the same species you do, but you have remote ancestors whom no one would call individuals of the species Homo sapiens. There simply isn't any definite line to draw between one species and the next, historically, and that was what was shocking about Darwin's idea of branching speciation from common ancestors, the only idea he presented visually in his book The Origin of Species.
For a book-length treatment of speciation as a scientific issue for the thoroughly curious, see Speciation by Jerry A. Coyne and H. Allen Orr,
which discusses mechanisms of speciation and how theories about those mechanisms are tested by biologists.
ONE MORE EDIT: The questioner in the other top-level comment is coming back with follow-up questions, which is good. The simple answer to the basic question is that biologists properly do NOT define a "species" by looking at DNA sequences, but by looking at reproductive behavior. For individuals who are dead and long gone, out inferences about reproductive behavior, and thus species definitions, are weak, and we shouldn't rely on a single DNA sequence to say a sample of old bones belongs to a different "species" from another sample of old bones until we gather other forms of evidence about the possibility of the individuals mating and having offspring with each other.
Svante Pääbo was the pioneer of doing DNA sequencing on samples of old human tissues (he started with Egyptian mummies when he was an Egyptologist). I heard a lecture by Pääbo at the 2008 Nobel Conference
https://gustavus.edu/events/nobelconference/2008/presenters....
and heard there from some of the other researchers who use DNA analysis to trace human origins.
One of the problems in the discipline so far, as you can well figure out for yourself by thinking about it, is that the sample sizes for defining what genome pattern is "modern human" are still remarkably small, especially when we get to full genome sequences of individuals with well determined geographic origin, and smaller still for Neanderthal sequences and smallest of all for Denisovan sequences. So much work still needs to be done to figure out what the usual range of free variation of genes that are not under strong selection pressure is in the hominin genome line.
Figuring out what genes actually do at the individual level in a genome as complicated as the human genome has turned out to be much harder than was first supposed (as mentioned in the submitted article), so while tracing lineages with gene markers that are not under strong selection pressure helps define geographical population groups with common founder populations, knowing even an individual's complete genome does little to show the traits of the individual. (The first human genome sequenced didn't even reveal that the sequenced individual has blue eyes, for example.)
Turkheimer, E. (2012). Genome wide association studies of behavior are social science. In K. S. Plaisance & T.A.C. Reydon (Eds.) Philosophy of Behavioral Biology (pp. 43-64). New York, NY: Springer.
http://people.virginia.edu/~ent3c/papers2/Turkheimer%20GWAS%...
"If the history of empirical psychology has taught researchers anything, it is that correlations between causally distant variables cannot be counted on to lead to coherent etiological models."
"Most Genetic Associations with General Intelligence Probably False Positives"
http://www.arts.cornell.edu/econ/dbenjamin/IQ-SNPs-PsychSci-...
Johnson, Penke, and Spinath (2011) "Understanding Heritability: What it is and What it is Not"
http://www.larspenke.eu/pdfs/Johnson_Penke_Spinath_2011_-_He...
"Presuming this is correct and, as Visscher and Keller appear to have implied, applicable to most behavioural traits, this indicates that literally thousands of genes are likely involved in each trait, with no single polymorphism having substantial effect, which is the quasi-in?nite model to which Munafo and Flint refer. Whatever would we as psychologists do with such information?"
AFTER EDIT: The interesting question raised by the top-level comment that was posted before this comment has received some good replies. The question, and the issue of defining a "species," reminds me of what Richard Dawkins points out about common descent. If you lined up in a line by order of descent with ALL of your ancestors, including the ancestors no longer living, as you looked at each individual in the line you could say "He is the son of the father standing next to him," going back as far as you want, even hundreds of millions of years, and yet at some point in the deep past the ancestors would not look anything at all like "humans," or indeed even like mammals or like tetrapods. At the individual level, OF COURSE you are part of the same species as your parents (by definition of "species"). Life on Earth today has a common ancestor, or at the very least a common set of one-celled ancestors, by general agreement of biologists, so all the species that have differentiated from one another over time can be traced to common ancestors of multiple species. You are undoubtedly related to your parents, who belong to the same species you do, but you have remote ancestors whom no one would call individuals of the species Homo sapiens. There simply isn't any definite line to draw between one species and the next, historically, and that was what was shocking about Darwin's idea of branching speciation from common ancestors, the only idea he presented visually in his book The Origin of Species.
For a book-length treatment of speciation as a scientific issue for the thoroughly curious, see Speciation by Jerry A. Coyne and H. Allen Orr,
http://www.amazon.com/Speciation-Jerry-A-Coyne/dp/0878930892
which discusses mechanisms of speciation and how theories about those mechanisms are tested by biologists.
ONE MORE EDIT: The questioner in the other top-level comment is coming back with follow-up questions, which is good. The simple answer to the basic question is that biologists properly do NOT define a "species" by looking at DNA sequences, but by looking at reproductive behavior. For individuals who are dead and long gone, out inferences about reproductive behavior, and thus species definitions, are weak, and we shouldn't rely on a single DNA sequence to say a sample of old bones belongs to a different "species" from another sample of old bones until we gather other forms of evidence about the possibility of the individuals mating and having offspring with each other.