Thursday, August 24, 2006
Dr. Luigi Luca Cavalli-Sforza is a professor of genetics at Stanford University. Dr. Cavalli-Sforza's magnum opus The History and Geography of Human Genes is a landmark of human historical population genetics, while his text coauthored with Walter F. Bodmer, The Genetics of Human Populations, is one of the most thorough introductions to the field of population genetics with an emphasis on our own species. Originally trained as a medical doctor, his work has spanned a range of fields from microbiology to theoretical anthropology. Recently Dr. Cavalli-Sforza's life has been the subject of a full length biography, A Genetic and Cultural Odyssey. One may view a list of representative publications here.
Below are his responses to "10 questions."
1) Can you tell us what R.A. Fisher was like as a teacher and a department head? It is sometimes said that true geniuses do not make for the best instructors or superiors.
As a teacher, his lectures were quite good. It may have helped that I knew at least some of the things of what he taught. As a superior, he was tolerant of his inferiors with whom he did not get along, but ignored them, and was absolutely loved by all the others, with whom he was very generous, and this was true of me also. When Kenneth Mather, one of his best students and with whom I had worked in advance learnt that Fisher had offered me a job, he said "in the first fortnight it will be decided whether you will be friend or not". I certainly became a very good friend of Fisher, and am a very good friend of all Fisher's last students whom I had a chance to meet and work with, like Anthony Edwards and Sir Walter Bodmer, both of whom have a profound attachment to him. Fisher was very impatient with bureaucracy (who isn't, unless he/she is very lucky?), especially that of the University of Cambridge. He supported two of his potential, excellent successors: Guido Pontecorvo and Francis Crick, but his advice was not paid any attention.
2) When I read "Consanguinity, Inbreeding, and Genetic Drift in Italy" I was struck by the relatively mild differences between the two geographic halves of the nation. Considering the historical legacy of Magna Graecia, Byzantine and Arab rule in the south, in contrast to the Gaulish, and later Germanic influence in the north, I was expecting more between-population differences. I suppose that gene flow across demes along lines of trade equilibrated allele frequences between regions (while high relief resulted in enforced endogamy). If you agree with my assessment, were you surprised by those particular results?
The only "genetic" data in the book regarding the country as a whole are from surnames, which have little time depth: at most one thousand years, and reflect the cultural more than the deep genetic background. Nervetheless they are highly correlated with it, but cannot be used for between countries comparisons. The cultural unity of Italy is older than a thousand years; more than three thousand years in the south and two thousand years for the north.
The truly genetic data can be found in The History and Geography of Human Genes, especially pages 292-293, showing that northern Italy is more similar to western and central Europe, the center to the northern Balkans, probably reflecting in part the middle Neolithic diffusion, and the south the southern Balkans, reflecting the earliest Neolithic diffusion and the Greek and Phonecian colonizations of the first millennium BC.
3) I suppose it can be said "The History and Geography of Human Genes" was your most influential book. The PC maps as well as the cladograms were mental candy for those of us who were interested in the narrative that it told about human demographic history. Nevertheless, I have noted a tendency amongst some to view the conclusions you offered as a sort of "proof text." As an example, the affinity between Southern Chinese and Southeast Asians, and between Northern Chinese and other Northern East Asians, have been the subject of much discussion on internet forums. Did you anticipate that many would view your inferences as indelible Truth?
Nothing is indelible, or will be indelible until data from a much wider selection of markers will be available. In this century I hope there will be a chance of examining very extensive samples of individuals for the whole genome and this knowledge will be much more indelible, as much as the representativeness of the samples tested. I doubt that I will see it, but this is the rule of life.
4) Moving to, in the interests of frankness, less influential books, in "A Genetic and Cultural Odyssey" Linda Stone & Paul F. Lurquin note the relative lack of response to "Cultural Transmission and Evolution" within the social sciences. You seem to chalk this up in part to the lack of comfort with mathematical methodologies within cultural anthropology. Over the past few years a small group of anthropologists, Peter Richerson, Robert Boyd and Joe Henrich seem to be continuing the attempt to model culture using the techniques that have been fortuitous in the biological sciences. Do you think that we are past the high tide of 'interpretative' anthropology and that a more explicitly hypothetical-deductive methodology may come to the fore?
I entirely agree that the average quality of anthropological research, especially of the cultural type, is kept extremely low by lack of statistical knowledge and of hypothetical deductive methodology. At the moment there is no indication that the majority of cultural anthropologists accept science - the most vocal of them still choose to deny that anthropology is science. They are certainly correct for what regards most of their work.
5) "The Genetics of Human Populations" was recently re-printed. In the foreword, you and Walter Bodmer acknowledge that 30 years is a long time in scientific history, but that your work does offer theoretical insights which have stood the test of time. I was interested to note, for example, on page 530 you offer an estimate of the number of loci which result in the normal human variation of skin color at about 4 using classical genetic methods. Recent comparative genomic work suggests that there are other loci besides the well known MC1R, with a recent gene, SLC24A5, accountable for 30% of the difference in complexion between Europeans and Africans. It seems that the estimate of 4 loci may not be very far off the truth. Do you believe that the "postgenomic era" will by and large vindicate the synthesis of population genetics and evolution which crystallized in the 1930s under R.A. Fisher, J.B.S. Haldane and Sewall Wright? Or, do you intuit from the most recent batch of findings that we are in for a data-driven "paradigm shift" which will result in a new theoretical framework?
I was relying on the opinion of Neil Risch that our book is still useful because it contains knowledge that is important and is ignored in more recent books. Everything changes, but I believe the old paradigm that mutation + nat.selection + drift + migration is sufficient to understand biological evolution, and the basic theorems are still the same, with few additions. Of course, the use of simulation - of which I gave probably the earliest example for the role of drift on blood group genetics in the Parma Valley in 1967 (cited in Consanguinity, Inbreeding etc.)- has given much greater power to theoretical analysis, but it can hardly give results of the same generality as the mathematical approach. When one goes over to cultural evolution or other examples of evolution of self reproducing systems one needs to consider mechanisms of hereditary transmission other than the (generalized) Mendelian one. Also in biology it will be necessary to consider the general importance of lateral transfer, which will bring a partial revolution, but in cultural evol. it is imperative to give to it much importance, especially in the internet era. Still, it is difficult to completely ignore the importance of parent-child transmission in cultural evolution, as well as the niche influence (Odell Smith, Laland and Feldman), when one considers (e.g.) some surprising permanence of pagan rites in Europe.
6) In "A Genetic and Cultural Odyssey" you seemed rather unrepentant about not explaining the details of Principal Component Analysis though you obviously recognize the importance of mathematics to your work. The historian of science Will Provine has asserted Sewall Wright's original work which coalesced into the Shifting Balance was simply not internalized on a mathematical level by most workers, including disseminators such as Ernst Mayr, and that this resulted in subsequent confusions about the importance of random genetic drift and gene-gene interactions. Obviously biology is a science that is focused on a particular subject, life, as opposed to being hewed to any one method. Nevertheless, do you perceive in the next generation of trained biologists the mathematical aptitude and inclination to tackle problems both analytically and computationally so as to extract more insight from the excess of data generated by modern sequencing techniques?
I adore principal components, as you may have noted, and I was the first who introduced them into genetics (1963, Proc.Intl.Genetics Congress of the Hague). I also showed with Piazza in 1976 (Theor. Pop. Biology, 8: 127-165) that when a simple evolutionary model of mutation and drift is correct, principal components and trees give the same result, in the sense that eigenvalues and vectors calculated in the spectral analysis of populations x gene frequencies matrices correspond to the meaning and order of the nodes found in trees. But I have always found it difficult to explain rigorously principal components to people who have little confidence with mathematics, and I decided to test the skills of Lurquin and Stone, hoping they would do it better than what I usually do. Of course, the usual indication that it is a method to reduce the number of dimensions gives some ideas, but without an appreciation of the amounts of variance explained it may give a false sense of security .
7) Question #3 hinted at the powerful social impact your work has had in reshaping how we view the natural history of our species. One of the most contentious issues of the 20th, and no doubt of the unfolding 21st century, is that of race. In 1972 Richard Lewontin offered his famous observation that 85% of the variation across human populations was within populations and 15% was between them. Regardless of whether this level of substructure is of note of not, your own work on migrations, admixtures and waves of advance depicts patterns of demographic and genetic interconnectedness, and so refutes typological conceptions of race. Nevertheless, recently A.W.F. Edwards, a fellow student of R.A. Fisher, has argued that Richard Lewontin's argument neglects the importance of differences of correlation structure across the genome between populations and focuses on variance only across a single locus. Edwards' argument about the informativeness of correlation structure, and therefore the statistical salience of between-population differences, was echoed by Richard Dawkins in his most recent book. Considering the social import of the question of interpopulational differences as well as the esoteric nature of the mathematical arguments, what do you believe the "take home" message of this should be for the general public?
Edwards and Lewontin are both right. Lewontin said that the between populations fraction of variance is very small in humans, and this is true, as it should be on the basis of present knowledge from archeology and genetics alike, that the human species is very young. It has in fact been shown later that it is one of the smallest among mammals. Lewontin probably hoped, for political reasons, that it is TRIVIALLY small, and he has never shown to my knowledge any interest for evolutionary trees, at least of humans, so he did not care about their reconstruction. In essence, Edwards has objected that it is NOT trivially small, because it is enough for reconstructing the tree of human evolution, as we did, and he is obviously right.
8) Your lab has recently been focusing on Y chromosomal lineages. It seems that this is a clear extension of your previous program of analyzing demographically informative loci to aid in reconstructing the natural history and behavioral ecology of our species. A graduate of your lab, Spencer Wells, has outlined the utility of the Y chromosome in his recent book "The Journey of Man." Nevertheless, others, such as Henry Harpending at the University of Utah have argued that functional loci which have been shaped by selection should be the object of greater interest, and that neutral loci can only tell us so much about the nature of our species. Of late, data emerging from the HapMap has hinted at powerful selective pressures upon our species within the past few thousand years. Some would argue that the high frequency of deleterious recessive diseases such as Cystic Fibrosis (frequencies too high to be maintained by mutation-selection balance) should have prepared us for this. What is your assessment of the alternative program which is focused on signatures of selection as opposed to lines of ancestry?
It is obvious that natural selection has played a major role in human evolution, although most statistical methods for detecting it are rather gross. Quantitatively, however, I am sure that drift has been very important - it will take more work on different types of genes to see how much, relative to selection. For an estimate (78%) based on microsatellites, see Sohini et al., nov. 2005 Proc. Nat.Acad.USA. 102 : 15942.
9) In "The Genetics of Human Populations" you allude in passing to intergroup selection (page 750), and seem to suggest it played a non-trivial role in the reshaping of the overall character of our species' genome. To be clear, if you do accept group level selection, would you expect its power (vis-a-vis individual level selection) to increase or decrease with the rise of complex mass societies and the diminishment of tribes & clans?
It is a good question. I suppose the answer is likely to be (tentatively) positive.
10) Over the past 6 years you have had a somewhat heated relationship with Bryan Sykes of Oxford in regards to the apportionment of European ancestry into "Neolithic" and "Paleolithic" quanta in the context of the model of Demic Diffusion. Moving past the details of this scientific relationship, which interested readers can explore in your aforementioned biography, there have been other recent extractions of ancient DNA of supposed Neolithic immigrants which have been used to "refute" the contribution of this population because of the lack of contemporary descended haplotypes in most of Europe. Others retort that neutral markers by their nature may be subject to substitution and that over generations most lineages become extinct. The technical details are clearly lost on the public, and yet there is also a great interest in this topic. Have you any advice for a potential science reporter who is attempting to cross and bridge this chasm?
Neutral markers are clearly subject to less rapid evolution than markers under directional selection. Markers under balancing selection are more stable than neutral markers. There is no estimate of the quantitative importance of balancing selection. A list of 91 genes showing selection at the genome level (Sabeti et al., 2002 if I remember the date correctly) show that about 25% show mostly balancing selection.
Archeology can be very useful for tracing history, clearly. But genetic analysis on fossil bones requires hard work, and with the investment of time and money that is possible today the numbers of individuals and markers on which archeological conclusions are based is too small to be considered respectable, especially when examining very small genetic differences.
Labels: 10 questions