Tip:
Highlight text to annotate it
X
Vence Bonham: I think it’s important to start with a couple
disclaimers. One, I’m not a geneticist. I’m a social scientist; a lawyer and health
services researcher by training. I spent nine years teaching medical students, and so I
come to the Genome Institute with a very different perspective of thinking about questions. And
so what I want to share with you is some questions that you may want to think about as you look
at studies, particularly studies that are talking about racial differences based on
genetics.
So one of the things that I want to share with you as a beginning, is, in 2003, at the
time of the completion of the Human Genome Project we developed a number of challenges
for the future of genomics research. And it’s quite interesting that over the four years
since the time of that article, that this challenge has continued to be one that I think
is extremely important to the field of genetics and the field of science, and will continue
to be. But I think it’s important to recognize that this is an issue that’s a challenge
and was identified as a challenge to society and as we think about these questions. So
the question of understanding the relationship between genomics, race, and ethnicity, and
the consequences of uncovering these relationships. So as we think about this question from a
societal perspective, I think it’s important that we think about both history as well as
the future.
And so I start with a context. And we can think about the last few weeks, we can think
about the last few years, but clearly the issue of race in America is something that’s
very important to people in our society and how people think about difference; how people
think and discriminate against individuals. Originally, when I was putting my slideshow
together I thought of having a picture from Katrina or a picture of Don Imus, or some
other kind of picture that was fairly immediate of how race always surfaces up in conversations
within our society. But race in science often does, and we can look to history as we think
about that question. So as we think about how we think about difference, how we think
about variation in America, becomes extremely important.
So I decided to propose questions that you may want to think about as you’re talking
to an investigator who has just reported on their study or if you’re thinking about
writing a specific story or a piece that may be on TV. What are some of the questions you
may always want to ask? And so I decided to present a few questions for you. And these
questions are both from my colleagues within the Genome Institute, but also colleagues
within the public health and the medical community, asking them, what would you ask if you wanted
to ask about studies related to human genetic variation and race?
So the first question: what was the basis for choosing the populations to study? How
did the study investigators categorize or define the study populations? Is the variation
peculiar to a specific group, race, or ethnicity, or only more common in it? And what are the
potential causes of the differences in outcomes beyond genetics? And what are the social implications
of misinterpretation of the data? And finally, what is the medical significance of the variation?
What does it mean for health or disease, or the “so what” question?
So what was the basis for choosing the population’s study? And I think this is really a challenging
and interesting question when we look at different populations across the globe. Is there an
issue of convenience or is there an issue of known differences with regards to disease
and disease prevalence? How do you think about that? How do you challenge the question of
why a specific population or group was identified for the specific study, is an important question.
And how did the study categorize or define the study population? And this is one of the
areas that I’m really interested in for my own research, is how people make a determination
with regards to how they’re going to describe the groups within their studies. So when you
think about that from the perspective of the United States particularly and the history
of how we’ve collected data around individuals and groups within the United States, there
has been a change over time. And so I just wanted to share with you some U.S. Census
data categories that we’ve used over the history of this country. And so, if you’ll
recall, back in 1790 we only collected really in two groups: free, white or other, and slave
was how we collected information in our country about groups and populations. But then if
you look at 1890 we collected groups by white, black, Mulatto, Chinese, Indian, Quadroon,
Octoroon, Japanese. Quadroon. Think about that and think about that related to genetic
studies. How do we understand? What are we trying to define? How are we trying to describe
people and difference? And then, more recently, you can see from our OMB categories. And here
I just wanted to highlight the italics where within the Census categories it clearly stated
that this is neither a scientific nor anthropological in nature.
And then finally, our current categories that we have for the collection of information
within our country. And you see the five racial categories and then the ethnic category, which
is Hispanic or not Hispanic. But what changed in 2000 was your could identify more than
one race, which I think is extremely important with regards to how we understand people and
groups within our country.
So another way to kind of describe this, is when you think about research studies and
studies where we’re looking at different populations. We really put people in bins,
and we don’t fit in bins. We’re not all in one bin. And that’s the difficulty and
the challenge that we have. But that’s really the only way you can really do the study.
And so the question is, why did you put them in that bin? Does it make sense that they’re
in that bin? Can you understand that bin? And is ancestry a better bin than our OMB
categories of race at times, or is it not? Or are there other ways to do this?
So the third question is can a specific ethnic group -- and I identify the HapMap populations,
but this could be another group that’s identified -- be generalized to all individuals within
a U.S. Census category of people. And so it really raises the challenge -- and this really
ties with Larry’s talk this morning -- is the question of our understanding of human
genetic variation and thinking about that and thinking about the level of diversity
that happens to be in the continent of Africa and how does that fit with our understanding
of thinking about different groups? And so many genetic studies have shown that there’s
clearly more variation within the continent, as only a small group of individuals left
the continent to cross through different parts of the globe. And so how do we understand
that, when we make differences related to skin color and identification?
So this is a slide from Lynn Jorde, a college of mine and a geneticist at the University
of Utah, where he does a nice job of really kind of articulating the issue of who’s
an African-American. And this is based on ancestral markers. So these are individuals
that self-identified as being African-American or black. And if you look at the two individuals
here, you see that one of the individuals has clearly a majority of their markers that
would be identified as African ancestral markers, and the other individual has a majority of
markers that would be identified as European markers.
I think we probably all know people like that; people who clearly come from various ancestral
backgrounds but they identify their identity, their social identity, which is so important
and a component of our lives, as one or another group and they make a distinction about that.
My own lab is doing work around understanding of how individuals who are clearly multi-racial
where one parent is European background and one parent is African background perceived
their identity and what happens within their clinical encounters around that identity.
So this question of how we think about individuals and describe individuals becomes extremely
important. But there clearly is a relationship between our self-identified race and genetics,
and it’s extremely important to understand that, that you can correlate, but it’s often
imprecise and the correlation of ancestry is clearly more informative. And so this is
a challenge, both for the researcher and how do you describe your groups, and use racial
information that can be very valuable social construct of race but also try to be very
clear in the ability to communicate when you’re talking about an individual versus a group
or population.
So, next question. What are potential causes of the differences and outcomes beyond genetics?
And Emily talked a little bit about this this morning. And some of the questions that were
raised is that we have to be careful and thoughtful. And, as a social scientist these are the issues
that are most important to me, is how do we think about the social determinants and what
that may make with regards to who has certain diseases; what happens with regards to disease;
differences in outcomes in morbidity and mortality; that the socioeconomic status, issues of transportation,
housing, access to service, education, income have so much impact on what happens to individuals
in our society? And so when we think about who gets diabetes, clearly the genetic component
is an extremely important component, but the cultural component, the diet, the issues of
access to general healthcare services can be also very important to those issues. So
how do we interpret and communicate that to the public, I think is a challenge for researchers
and is likely a challenge for you.
So, I wanted to identify one study that was published a couple years ago that you may
or may not be familiar with, which was published by the Institute of Madison called “The
Unequal Treatment.” And it’s really been a profound study in our understanding of health
disparities and really raising issues and providing a consensus of data that there’s
difference in treatment that people receive within our country that is related to the
race and ethnicity, that has nothing to do with the genetic basis but has to do with
how they’re treated within a system, can have equal insurance, but where there’s
differences in how they’re treated, and what does that mean related to health and
illness, or may not mean, related to health and illness? But the importance of understanding
that the society and the influences on that with regards to what happens with individuals
and their health is extremely important.
So I want to share with you a study that was published by Richard Cooper and a group of
researchers that I think does a nice job by just kind of raising the case study of why
we all have to be thoughtful. And this was the study related to blood pressure in populations
of European versus African descent. And specifically, he was looking at the question of hypertension,
which is often identified as African-Americans and individuals who are self-identified as
black as being greater risk generally throughout populations across the globe for hypertension.
And so he wanted to look at that from a perspective of whether trying to tease out what may be
the differences in different populations. And this figure from the article identifies
a number of different population groups. So you’ll see here where U.S. blacks are a
little over 40 percent prevalence of hypertension, and Nigerians a little bit over 10 percent,
and Jamaicans about 25 percent, but you see the highest is from Germany. And so the question
of challenges of understanding environment and diet and culture become extremely important
in this question. And so it’s really a question to challenge that we need to think about all
of the issues with regards to how we understand disease indifference within our society. And
so that this article is just an example of challenging us to make sure that we don’t
over-focus on the genetics, but that we also look at the environmental and the social issues
that are extremely important.
So the next question is, what are the implications of misinterpretation of the data? And I think
that there are a number of examples from history that we can look to, but I just want to give
one example that I want to share with you. And this is from Ebony Magazine in 1966. And
I just want to highlight this article because it described Sickle Cell Disease and it talked
about Sickle Cell Disease as a *** disease. And clearly within the United States the prevalence
of individuals who have Sickle Cell Disease would be identified as black or African-American.
But if you think now our knowledge that we clearly have across the globe, if you’re
in Greece or in Saudi Arabia or other parts of areas of the world where there was a high
Malaria possibility at one time, there are a number of individuals who also have Sickle
Cell Disease or are Sickle Cell carriers. And it raises those questions. So this is
from JAMA in 1947, and I know this is from 1947 and we -- [unintelligible]-- and, you
know, it’s always nice to have 20-20 hindsight and to be able to look back, and I think that’s
a challenge that we all have as we move forward. But in JAMA 1947, the most significant feature
of Sickle Cell Anemia is the fact that it’s apparently the only known disease that is
completely confined to a single race. But clearly that’s not true. And so how we think
about this question -- and I think Sickle Cell Disease, even though it’s a single
gene disorder, is a great case study of issues of misinformation, of discrimination against
a group, a stigma against individuals who may be at risk for something but not have
a disease that we can learn from when we think about society, the United States, and race,
genetics, and disease.
So the next question is, is the variation peculiar to a specific group, or only more
common in it? And I think this is an extremely important question as we talk about differences
in groups related to potential genetic variance and how does the investigator to you as a
reporter when you ask that question and are able to interpret their results? So, how is
variation distributed between populations? Larry talked about this earlier. But again,
this slide from Lynn Jorde does a really nice job to show that the green bar that most of
the variation is within a population group. And so the differences between populations
is really very small, when you look at the variation that occurs within populations and
between populations.
And what is the medical significance of the variation? What does it mean for health and
disease? It’s the so what question so often for clinical researchers, you know, what impact
is this going to have on the health of my patients? And I just wanted to kind of give
you a context on a couple of things. One is an issue that I think has been a milestone,
but probably a milestone for the wrong reasons. And it gets talked about a lot in genetic
and genomic conferences and it really wasn’t a genetic or genomic study whatsoever, even
though it raised some questions about potential genetic variation. And that’s the issue
of the first drug that’s been approved by the United States for a specific racial group,
which was BiDil. And this is the indication and usage direction where it says, “BiDil
is indicated for treatment of heart failure as an adjunct to standard therapy in self-identified
black patients to improve their survival.”
But the challenge is who is self-identified black, and how do you understand that, and
what does that mean from a clinical care perspective? And really, are there other things that are
going on that we need to learn and understand that may make BiDil very valuable for a lot
of populations that would not identify as self-identified black or for individuals that
clearly are self-identified as black where it has no benefit to? And the question that
I’m just really raising for you is that from a clinical perspective, we have to always
be thoughtful of the challenge of what does this really mean to improve the health of
people in a society; the people that may read your articles, the people that may see your
T.V. pieces, or the scientists that interpret others’ studies? And so it becomes important
that we’re thoughtful about these questions.
So I’m going to end on a couple -- I’m going to share a little data from some work
that I’m doing, research that I’m doing, about how physicians think about these issues.
What do doctors think when we think about human genetic variation and genetics and race?
And I’m just going to share with you two quotes from some focus groups that we’ve
done with general internists across the country that I just think are nice illustrative points
of the issue. So the first quote is, “So I think we clearly know that biologically
there are different chemicals and different functional processes going on in African-Americans
versus white patients. You have to take that into account. They’re going to respond better
to one drug more than another, perhaps.” This is a physician in Detroit, Michigan.
Second one. “Race is a social construct. It’s useful, given the historic context.
I would not let anybody not refer to me as a black woman, however, I actually do believe
that there’s no biological basis for that and science proves me out.”
So we are entering into an exciting period of better understanding the genetic components
of disease. And then how does that translate both to society and the public and how does
it translate into clinical care becomes extremely important.
And I just want to end with a quote from a piece that Francis Crick did in 2003 as part
of the celebration of the completion of the Human Genome Project and the anniversary of
their identification of the structure of DNA. And Francis Crick made this statement. It
was a very powerful but short interview that he did. He was ill at that time. But he made
this statement that I think is important for all of us to remember. “In biology, a new
world is expanding in front of our eyes. May we learn to use it wisely.” Thank you.
Questions?
Larry Thompson: We’ll start with questions.
Vence Bonham: I got everybody quiet.
[laughter]
Larry Thompson: So actually, so I mean I actually have a question.
Vence Bonham: Okay.
Larry Thompson: So I want to start just sort of -- so biologically,
at the biological level of the gene, is there a distinction between blacks and whites?
Vence Bonham: So, no. So I guess the issue is that what
is black, what is white, and that there’s clearly information and data that supports
ancestral differences. And so you can see human genetic variation across the globe that
is ancestrally linked, that may have some links with also skin color and other features
and traits that you identify individuals from a social construct of race. But our social
constructs of race are that, they’re social constructs, but they may have valuable information
to describe different ancestral backgrounds that are important to understand human genetic
variation.
Larry Thompson: And so as we go forward, you know, we’re
going to be hearing about studies. We’ve heard the studies on obesity. We’re going
to hear some studies coming out shortly about height. And from what I gather when -- this
is unreported information at this point, so this is just for your information -- the distribution
of that trait about height, when you look at the HapMap populations, is different. So
it’s different and it’s not as prevalent in Asians who are shorter than Europeans who
are shorter than the Eurobans, who are the North African samples, or the Western Africa
sample. So I mean what are we supposed to do? How do we interpret that? What are we
supposed to do with that information in terms of discussing how these traits that are not
diseases vary across the populations across the face of the planet?
Vence Bonham: I think that’s a perfect example. And if
you think about the continent of Africa and the HapMap populations, Euroba, which is from
the Nigeria area, so the West Coast, but if you think about the continent of Africa and
the height of individuals from the continent, it changes dramatically. And so there are
certain tribes and ethnic groups from Southern Africa that are very, very short and from
Eastern Africa that’s very, very tall. I mean that’s a perfect example that you then
couldn’t say that individuals who are “African-American,” or who are black, are going to be taller,
or shorter, or this. So I think that’s a great example to show the variation that occurs
within different areas of the globe and within one continent. Yes?
Female Speaker: Years ago I read and heard of a theory that
it, partly explaining hypertension in American blacks. And I think it went like, the blacks
and the slave ships who could, you know, many, many died, but the ones who could retain salt
the best survived. If I’ve got that right.
Vence Bonham: Mm-hmm.
Female Speaker: And they brought those genes to this country
and that that explains why American blacks have more hypertension than African blacks.
Is that -- was it ever true and is it still true that there’s --
Vence Bonham: So I’m going to look to my colleague, Doctor
Brody, to give a perspective on this.
Dr. Larry Brody: This is a persistent and pleasing story, although
I have not seen any data that would suggest that it’s true. The other issue is that
to change allele frequencies, to change frequencies in population, takes a lot, a lot of selection.
And as horrible as the passage was, it may not have been a strong enough selective force
to actually change the frequencies in populations. If you think about the passage over the slave
trade, it occurred over hundreds of years drawing from different regions. So I haven’t
seen published information that says this is what happened and this is why this might
be true. And as you saw, the range of blood pressure varied and is almost a continuum,
almost without regard to continent. So if there are some documented differences, I don’t
know about it in the genes. And clearly the environment changes quite radically across
Vence’s profile from Nigeria all the way to Germany. I don’t know if Elaine or Francis
has ever seen any scientific publication that said that this might be the --
Larry Thompson: No, because at the moment we don’t have
any good example of genetic variations that have been validated as showing a strong association
or even a weak association with hypertension. Those have been pretty frustrating to come
by. And at such time as such variants appear, I suppose this could finally be a testable
hypothesis by looking at the gene pool in African-Americans compared to West Africa
where the slave trade primarily was conducted and you could see if there was any shift.
Like you, I’m quite skeptical that that will be found, but it’s testable.
Dr. Elaine Ostrander: But why wouldn’t it have been enough time?
You know, if there were a strong pull for a certain genetic type of individual to survive
and the others not, why wouldn’t that be enough time?
Dr. Larry Brody: It actually comes down to the math of selection
and replacing genes, or different flavors of genes. In order to get total replacement
it either takes a long, long time of drifting slowly, so little bits of selection over a
long, long time, or massive amounts of selection over a short time. And short time in this
case would be hundreds of years. In order to get full replacement so that you had a
big shift in the frequency, you may have to have 90, 95 percent mortality. And the mortality
was awful but it wasn’t that high. Once you have people over here and then you no
longer have the selective force it’s really easy to have all of the alleles maintain a
frequency. It also fits with the myth of how you could get rid of recessive disease by
eliminating the people who are affected. There are lots of recessive diseases that are lethal
and people don’t reproduce. But there’s this huge pool of people who are heterozygotes
who aren’t affected who continuously replenish that disease population. So in large populations,
allele frequencies don’t change radically unless selection is occurring over a long,
long time or selection is extremely intense. And there’s one example that’s been published
over the last year in the ability to metabolize lactose, and that’s a story where there’s
an example in the human population. It looks like some populations have acquired the ability
to use milk for nutrition as adults, which not everyone can do, and the signature on
the genome looks like that that trait is a recently evolved, recently selected for trade.
But those are relatively few and far between, those examples. One of the things that we’re
trying to do is use this HapMap-type data and whole genome association data to look
for other signatures of selection in the genome, but there’s not a lot of them at this point.
Sickle-cell is another one where, or anything that affects hemoglobin in malaria-ridden
areas is another one. But we’d run out of examples very quickly if we went through them.
Dr. Elaine Ostrander: You know, just to put another perspective
on that. We do a lot of modeling of population genetic selection in the 150 dog breeds, because
we do see excessive of particular diseases in one breed or another. So if you’re a
Scottish Terrier, your chances of getting bladder cancer are 30fold higher than if you’re
any other breed of dog. And so we look at exactly these sort of selective pressures
and selective sweeps and what breeders were selecting for and what got dragged along with.
And the pressure that you have to put on to really change the allele distribution is absolutely
dramatic. When we see something actually change we often see a population that for instance
bottlenecked during World War II down to maybe 20 dogs, of which maybe only three or four
were stud dogs that were still breeding, contributing to the gene pool. We know of examples in the
Liambergers where they bottlenecked down to about eight dogs. And so the pressure that
that population drops enormously, the number of individuals in the population contributing
to the gene pool drops enormously, and then there’s a very, very strong selective pressure
applied to those remaining individuals, then you can start to see changes in allele frequency
in a population. But is almost takes the constellation of those three events to see some meaningful
change. It has to be really quite dramatic.
Female Speaker: I have a question for that regarding Ireland;
the, first lighthearted. When will Irish Catholic make it into the OPM category for HapMap?
Because they didn’t like that when I wrote that in for my SATs many years ago. But regarding
folklore in Ireland growing up, we had always grown up hearing that because of the famine
and because of the famine years and because of forced immigration and starvation and mortality
that the tallest and the strongest were selected to go out on the coffin ships to go and live
in Australia and New Zealand and America; that the families chose those who were going
to make it. So when you look at the population now in the Republic of Ireland, the stature,
the physical stature of the men is less tall and big than today’s NFL linebacker. So
I’ve always thought it was an interesting question and certainly never has been, as
far as I know, explored scientifically, but I think it would be an interesting area for
study.
Vence Bonham: Well the whole issue with the salt retention
and the migration, I mean if you follow the articles and really look at the articles you’ll
see that this is a hypothesis, that there’s been no empirical data. But I know that recently
there was a popular TV show that had a piece on this, and so this is one of the challenges
of how do we understand where the science is today, and how we understand some of these
differences.
Dr. Larry Brody: Let me address the Irish issue a little bit.
I collaborate with the folks in Ireland. Spent a fair bit of time talking to them and visiting,
and so I’ve picked up a little bit of the famine statistics. And it’s another case
where the famine, again, was awful and there was a huge efflux from the population, but
the numbers again didn’t have the selective pressure that we need to actually put allele
frequencies. And even if they did, even if you selected all of the tallest -- if you
remember the slide I showed about the distribution in height -- a generation of breeding, for
lack of a better term, will essentially have that sharp regress to the mean because we’re
talking about lots of different genes and difference combinations. And so, on average,
tall parents tend to have children that are more similar to the average than they are
to themselves, to a certain degree, although there’s a strong genetic component. The
population of Ireland, you know, is still not what it was during the famine, but the
number of Irish individuals in the world is roughly 10 times the population of Ireland.
And my guess is that you won’t find significant differences between the 40 million Irish that
are outside of Ireland and the five or six million that are there.
Male Speaker: In forensic medicine, how is it that we’re
able to identify the so-called race of the victim? Are we latching onto so-called characteristics
of curly hair or -- and do we even know the genome that well that we can identify the
gene responsible for this or that?
Vence Bonham: Francis, do you want to start that one out?
[laughter]
Dr. Francis Collins: So, as Vence told you, most of genetic variation
is shared amongst groups, but the frequency of a particular variant may be different depending
on where your ancestors came from. Remember Larry’s picture of those colored circles
migrating out of Africa and particular variants therefore turning out to be more frequent
in Asia or Europe than they might be across the continent of Africa because of a founder
effect or just drift, which is another term geneticists use for the fact that the frequency
will, especially in a smaller population, not stay constant just by random sort of stegastic
change. And if you look across all 10 millions snips in the genome, and you look carefully,
you can find some where the frequencies of a variant are really quite different, depending
on whether that person’s ancestors came from Europe or Asia or, in the case of HapMap,
the Euroba and West Africa. And that means that if you simply sample those that are particularly
ancestrally relevant with a mere 20 or 30 of those snips, you can make a prediction
about a DNA sample, whether that person’s ancestors came from one of those three parts
of the globe. And you’ll get it right as long as you’re not looking at Tiger Woods.
Larry Thompson: Yea, I was going to say the exact same thing
and just stumble on that person, yea.
Dr. Francis Collins: Right. So if you’re given a limited number
of options, you should be able to make a pretty good prediction about ancestral geographic
origins. To say nothing about whether that is a predictor of race, because race is such
a murky baggage-laden term, that in this instance I think one should be careful about what you’re
determining and what you’re not. You’re really determining ancestral geography of
that person’s predecessors. And certainly that is being used in some instances in forensics
to try to take a guess at what the origins might be of somebody who has left a DNA sample
at the scene of a crime. Even if you don’t have a match, maybe it’ll help the authorities
in their abilities to try to prioritize who might be the culprit. And there are at least
one celebrate or not celebrated, depending on your perspective, case in Louisiana where
that was done.
That’s quite controversial, because, certainly, given the fact that many people don’t have
all of their ancestors from just one part of the world, and certainly when you look
at the group that we call, for instance, Hispanics or Latinos, there’s a very different kind
of ancestral history if you happen to be from Puerto Rico, as opposed to Mexican-American,
as opposed to somebody who grew up in the country of Spain, and yet we would call all
of those Hispanics. So maybe this tells more than it should. Of course, this has turned
into quite an industry of how people can send in a DNA sample and be told where their origins
came from, right down to, in some instances, the village in Africa where your ancestors
came from, which is quite fanciful since, of course, Africa has not been static over
the course of all this time, either. And so some of that is people actually making bucks
on making predictions that are really not well supported, at least not to that level
of precision.
One of the things we are anticipating though is that as we get more and more information
about how genetic variation not only correlates with ancestral origin, but actually we identify
variations that are associated with appearance. And you heard about one from Elaine; this
gene that’s associated with skin color. It’s not unimaginable that in the next five
or 10 years we’ll also identify genes that play a role in facial structure, cranial facial
features, and hair texture and hair color and eye color will, if not already getting
glimpsed, get better and better so that a sketch of the individual derived solely from
their DNA sequence might get better and better. And then that really is going to, I think,
turn things upside-down in terms of forensics and in terms of privacy issues about who has
access to your DNA and what do they do with it?