Anthropology, Genetic Diversity, and Ethics 
A workshop at the Center for Twentieth Century Studies 
University of Wisconsin-Milwaukee  
Lynn Jorde
[Participant Information]

Well, thanks Jonathan.  I ... in fact, I was pleasantly surprised to learn that our research has been successful!  [laughter]  And I would like to talk with you a little bit about -- actually, mostly about the science, some of our findings, and then toward the end move on to some of the reactions to those findings, and Professor Naidu will talk further about the collaboration and about some of the reactions to our findings in India. 

But I'd like to start by addressing, I think, a more general issue, as a scientist.  Why are we interested in studying human genetic variation?  What is it that motivates us?  And I can think of several reasons why we do what we do.  One does involve the biomedical applications of genetic variation, studying genetic variation in populations, understanding the reasons for the variation that we see in the distribution and prevalence of genetic diseases.  Why is cystic fibrosis so common in northern European populations?  Why is sickle cell disease so common in African populations?  We are understanding better and better the mechanisms responsible for that variation, the evolutionary forces contributing to that variation.  So I think that's important in terms of understanding the genesis of those diseases, and in some cases, can contribute to our understanding of the etiology of diseases.  Studying genetic variation is important in terms of understanding our reactions to environmental agents.  One of the most interesting of these is reactions to prescription drugs.  According to an article published in the Journal of the American Medical Association last year, over a hundred thousand Americans die each year as a result of adverse reactions to prescription drugs.  One of the things that we are interested in is target -- or understanding how genetic variation contributes to those adverse reactions, which it clearly does, in many cases.  So there are enormous health benefits right there from understanding genetic variation in human populations. 

The second area, and one that I've been involved with now for a few years, is in the forensic applications of human genetic variation.  DNA profiling has been used in thousands of court cases now, both to successfully prosecute guilty individuals, and also -- just as importantly -- to exonerate innocent individuals.  One of the reasons that that evidence, when used properly, is successful, is that we have come to understand the nature of variation in human populations.  By understanding that variation and how it is distributed, we can better estimate, match probabilities, and come up with reasonable estimates that are admissible in the courtroom. 

And finally, an area that many of us are interested in is the elucidation of population history.  Our, every major event that affects human populations -- migrations, bottlenecks, selective events -- leaves an imprint on our genome.  And it's a fascinating exercise to look at our genome, to look at genetic variation, to deduce those events.  So what I'll be talking about today is sort of a test case, in which we've used a known, established mating system to test how well our predictions of genetic variation hold up when we actually look at the results of our genetic analysis.  Have I got the slides on?  You've got the focus, okay. 

So here are some of the research questions that we've been addressing in our studies in south India.  First of all, is there significant between-population genetic variation, and we've been looking thus far primarily at caste variation, and what is the pattern of between-caste genetic distances?  What are the likely origins of south Indian castes and tribal populations?  And finally, is there evidence of a sex-biased gene flow between castes?  And this is essentially the system that we were testing with mitochondrial and Y-chromosome genetic markers, and I'll talk a little bit more about that as we go along. 
Now, we can summarize -- and this is a very, very oversimplified summary of major historical events in India, but it gives us at least a framework from which to proceed -- Paleolithic settlement of probable African origins, migrations of proto-Dravidian speakers from the Fertile Crescent area about nine or ten thousand years ago.  And then most recently, a third major event, migrations of Indo-European-speaking, so-called Aryans from West Asia, about 3500 years ago, and it was these individuals who established, who are thought to have established, the caste system.  And I'll be referring to these waves of migration as we go through the talk. 

Now the Hindu marriage patterns prescribed by the caste system are of significant interest, and have been for many decades, to a variety of researchers.  Primarily, marriages occur between individuals of equal caste status, but men from higher varna or groups of castes may occasionally marry women from lower varna or groups of castes, but the opposite is very, very strongly discouraged.  So basically, we have a system in which males very seldom change caste, but females occasionally do, and most typically when that happens, they would change to a caste of slightly higher status or rank.  So that provides us with a mating system, one that's been going on for more than a hundred generations, that predicts certain kinds of patterns of genetic variation.  So what we're interested in is seeing whether those historical patterns are indeed reflected in the genetic systems we're looking at. 

The study populations that we have used thus far -- these numbers are actually a little bit old, but from the upper, so-called upper castes, about 60 individuals, middle, 160, and lower 73, and those are further subdivided into these groups shown here, so I won't go into that in detail.  Our primary comparisons have actually been between these three groups of castes, and these are all males [who we've typed?].  This shows the location.  Actually [inaudible place name] is down in this part of India, south India, and then we've also collected a number of samples from throughout the region, but I won't really be talking about those collections today. 

Now the genetic systems we've looked at -- I won't go into great detail about this, but primarily we've looked at sequence variation in the mitochondrial DNA.  This is DNA that you inherit only from your mother, so it gives us an idea of the maternal lineage of a population, and we can look at differences in mitochondrial DNA to deduce the degree of relationship, the degree of genetic similarity, between individuals in this maternal lineage.  We've also looked at Y-chromosome variation, and of course that gives us the other side of the picture.  The male, the paternally inherited lineage.  In addition to these Y systems, we've now, we're now in the middle of or nearly done with typing 30 systems in the autosomes, but again I won't be referring to those today. 

So if we look at variation among the cast populations, this statistic tells us what percentage of the genetic variation is the result of subdivision into castes.  So if we look at the mitochondrial DNA, the maternal lineage, we see that about 1.5% of the variation exists between castes.  The vast majority exists within castes, but about 1.5% is seen between castes.  If we look at the Y chromosome systems, about 10% of the variation that exists is observed between castes, 90% within.  So one important point here, of course, is that the vast majority of variation occurs within castes, rather than between.  But something I want to stress here is that 6 or 7 times more variation is seen for the Y chromosome polymorphisms, the Y chromosome systems, than for the mitochondrial systems.  That will be an important point, one we'll come back to later. 

If we look at genetic distances between caste groups for our mitochondrial systems, our maternally inherited systems, what we've done is to color the different groups, the upper castes here in green, middle castes in blue, and lower castes in yellow.  And one of the things that we see is that there's a fairly good grouping by caste, by caste ranking, when we look at differences in mitochondrial variation.  In other words, what this tree tells us, it tells us how closely related these individual -- these populations are, in terms of their mitochondrial DNA, their maternally inherited DNA.  And we're seeing a clustering of populations according to caste ranking, so those castes that are more similar in traditional ranking are more similar in terms of their mitochondrial DNA.  We can compare that to Y chromosome genetic distances, and now we see that there's really very little association between caste ranking and genetic distance.  In other words, caste ranking has no relationship, really, with the degree of paternal similarity, the degree of similarity in that paternally inherited lineage.  And if we look at the... 

So that result, I think, is a very interesting one, because it's exactly what we would predict, based on the known history of the caste system.  Females do switch castes occasionally; they can move up.  And so castes that are more similar in rank are going to be genetically more similar, just like populations that are geographically closer to one another tend to be more similar to one another, because they exchange mates.  So this laddering effect is what we would expect for mitochondrial DNA.  Now, for Y chromosome DNA, for paternally inherited DNA, because that DNA, because males seldom switch castes, we don't see that effect.  They stay where they are, and the variation that exists is largely due to random effects: genetic drift, mutation, and so forth.  So that corresponds to the model shown here, in which there is gene flow between castes, primarily for females, giving this sort of stepwise effect in genetic distance, but not for males.  So it's a nice, essentially, verification or demonstration that indeed, the history of this population, the history of the social mating patterns are observable in the genetic variation that we're looking at. 

Now, we can also look at -- and this is some fairly new work that hasn't yet been published -- mitochondrial DNA, genetic distances between various caste groups and continental populations.  And one of the things we see here is that all of the castes are most similar to groups of Asian populations.  And that's expected, given the origins of the proto-Dravidian populations, so that we see much, much smaller distances between Asian populations and each of the caste groups for the mitochondrial DNA.  When we look at Y chromosome DNA, we again see a somewhat different pattern.  For the upper castes, they actually have slightly greater genetic affinity (smaller distances) for European populations than for any other population.  The middle castes, a little bit smaller distance between them and Europeans, and the lower castes instead have the greatest affinity to Asian populations.  Now, if we think back about the history of the population, the most recent wave, and the one that is the most recent wave of migrants, the one that instituted the caste system is thought to have come from West Asia, Eurasia, and those individuals would likely have been more similar to individuals from Europe.  And as those individuals who began the caste system, they also likely appropriated the highest positions in the system.  So, a very interesting historical insight, again consistent with some historical hypotheses that the invaders who came in about 3500 years ago, established the system, and primarily who were male, so we see the Y chromosome versus mitochondrial difference, we can still see that signature in today's genes. 

So now, there have been a number of reactions to the publication of our first results, showing the difference in mitochondrial and Y chromosome variation in the caste groups.  And I'll just touch on this area, Professor Naidu will talk more about it.  Concerns haven't been addressed, such as this one here, "Social scientists worry that it won't be long before concepts like 'My DNA is better than yours' begin to take root," and I think this is an appropriate concern, and any time we're looking at variation between populations, [it's] something that we need to be aware of.  But it's also encouraging that in that same publication, the authors interpreted our results [and?] showing that this study has found that genes reflect social patterns, but the reverse is certainly not true: that their social status could not be nestling in their genes.  So we were pleased to see that there was a fairly accurate interpretation of the results, and certainly one that we have tried to encourage.  And finally, in this publication, they envisioned a scenario where genetics as a study becomes a partner of social sciences, and I think that 's an encouraging attitude, certainly one that we want to foster and perpetuate. 

So to summarize what we have found so far, I'm encouraged scientifically at our results, that show that indeed, history leaves a mark on the genome.  And in this case, where we have a historical system, a pattern that we could test specifically, we saw that our predictions were largely verified.  So that's encouraging for studies in which genetic variation is being used to deduce population history; it tells us that indeed, population history leaves a mark on the genome, and one that we can study and interpret.  And the secondary reactions to the study, we have spent, particularly Mike Bamshad has spent a lot of time talking with journalists from India to try to be very certain that our results are interpreted accurately.  Professor Naidu has also spent a lot of time doing that.  I think that for those of us who do studies of genetic variation in populations, that's a very important part of what we do, [an] important part of our responsibility, to be sure that the results are interpreted accurately.  Thanks very much.