First, I'd like to express my thanks to everybody who organized this
and did so much work to do this, and to everybody who's here, because we
don't get anywhere unless we do this. So I'm going to talk maybe
a little bit differently than some of the other speakers who spoke before
me, because I want to talk a little bit about the data behind what I do,
just to get sort of a background to start with. So -- could somebody
flip the switch there, turn the thing on? Thanks. Okay, that's
me. When you're dealing with ancient DNA, it's a little bit of a
special situation, because although when an individual is alive, their
body repairs DNA, [and] the damage caused by oxidation and hydrolysis and
radiation and chemicals, once that individual is dead, those processes
don't continue, and so DNA degrades over time. It gets broken, the
molecules get changed, and essentially, over enough time, DNA disintegrates.
And this leads to special problems, and it causes most of us to work with
a special kind of DNA called mitochondrial DNA. This is a bad slide
of a cell, and most of our DNA, billions of bases, is found in the nucleus
of the cell. We only have two copies, one from Mom, one from Dad,
but about 16,000 bases of our DNA are found in organelles called mitochondria.
They're what helps the cell produce energy, and there's 700 or 800 of them
per cell, so there are many more copies of mitochondrial DNA in a cell
than nuclear DNA. And that means that if you're working with ancient
DNA, you're much more likely to find intact mitochondrial DNA, a few copies
of it, than intact nuclear DNA. So that really limits a lot what
you can do with ancient DNA. All of the procedures and protocols
that we work with with modern DNA are not available to us. mitochondrial
DNA also has other advantages. It has a very high rate of mutation.
This means two individuals or two groups are going to have many more differences
in the mitochondrial DNA than they are in the nuclear DNA. And this
means mitochondrial DNA is appropriate mostly for questions of recent evolutionary
history, questions within a species' history. So that again limits
what we can do with ancient DNA. In addition, it's only maternally
inherited. That means a mother passes on her mitochondrial DNA to
all of her children; the father passes on none of his. And her daughters
will pass it on, but her sons will not. This makes mitochondrial
DNA a little easier to model and understand, but it also means that it
generally traces what's going on with women. What's going on with
men may be different.
Okay. DNA, of course, is made of polypeptide chains held together
in a double helix, with four bases, A, G, C, and T. It's the order
of those bases that hold the information of the DNA. And with ancient
DNA, what we do is we amplify the DNA. We target a specific area
of the DNA, and we amplify copies of it. Because even though mitochondrial
DNA has many more copies than nuclear, it's still very little DNA.
And so we use essentially the same process that we use in our own bodies
to copy DNA, only we take it out into the lab, and from one copy we can
make two, from two copies we can make four, from four eight, and it only
takes a few minutes to do each copying. So, from a few copies you
can, in a few hours, end up with millions or even billions of copies of
a very small fragment of the DNA. When we have this copy we can analyze
the mitochondrial DNA in a couple of ways. A coarse level of resolution
is what's called RFLP, or restriction fragment length polymorphism.
And this is when you take enzymes that recognize a very specific set of
DNA bases. For example, here, Alu I recognizes AGCT. So if
it goes along the DNA chain, which is a double chain, and sees AGCT, it
will cut the DNA in half right there. And we have surveyed many different
populations, many different individuals, the whole mitochondrial genome,
for where the most common cut sites are, which means that we can look at
those specific sites and see if an individual's DNA cuts there or doesn't
cut there. And the presence of a cut site that, in the majority of
humans, isn't there, or the absence of a cut site that in the majority
of humans is there, means there was a mutation in this small region of
DNA, so that it doesn't get recognized anymore. And when we look
at the pattern of these cuts, it turns out that mitochondrial DNAs can
be classed into groups that share certain cuts. For example, this
mitochondrial genome is a circular genome, rather than one big long chain.
There are four sites here that are sites that are commonly found among
Asian and Native American individuals. And -- here they are again
-- these divide these individuals up into groups who share mitochondrial
DNA, that's inherited from the maternal lineage, that is more closely related
to each other, because they share these mutations, and they can trace their
ancestry back to a common maternal ancestor. That's true of this
-- these are called haplogroups -- of D, and then these two haplogroups
can trace themselves back to an even earlier maternal ancestor, and that's
true for the whole human species. You go back far enough, and we
trace ourselves all back to one female human ancestor.
The other way we can look at mitochondrial DNA is to look at the actual
sequence of a fragment of DNA. We tend to look... this is again that
circular region, and it's got all of the genes -- there are 37 genes on
mitochondrial DNA -- labeled, and then there's a region of DNA that doesn't
code for a gene. It's usually called the D-loop or the control region,
there's a couple of names for it. And it's even more variable than
the rest of the DNA in mitochondrial DNA, and that makes it useful for
identifying differences between individuals. Most individuals that
are not closely maternally related are going to have differences here.
And it turns out that some of these differences can be correlated with
these haplogroups that we defined with the restriction enzymes, these RFLPs.
For example, these are individuals from the Pacific. This isn't in
very good focus. And you can see, these are the bases that vary in
these individuals in a D loop, and there are some obvious patterns here
-- depending on the focus. There's a bunch of individuals who share
these ACTs, and in fact they also share some markers out here, and those
also correlate with other restriction sites that make them a haplogroup,
which is very originally named Haplogroup 2. But in addition to that,
these individuals have other mutations that make them different from each
other. And sometimes they can be found in subgroups that are geographically
located; sometimes they don't seem to be.
So, using these two tools with mitochondrial DNA, we can help answer
questions that are of interest to many fields: anthropology, archaeology,
linguistics, history, conservation biology, evolutionary genetics, forensic
sciences -- I'm doing the promise part of my talk right now [laughter]
-- and some of the examples that I'm familiar with as an anthropologist
involve looking at kinship among ancient individuals, and this can help
us understand social structure, marriage patterns, burial customs, in these
ancient groups. In addition, you can look at nonhuman ancient DNA
and usually identify the genus and sometimes the species of ancient animals
found at the site. This helps us understand ancient dietary patterns,
hunting patterns, the ecology at the time the site was inhabited, and also
when and how animals and plants were domesticated in different areas.
We can also detect the DNA of ancient diseases. One common example
is what was the distribution of tuberculosis in prehistory, and that has
implications for its treatment today.
Also, we're just now beginning to be able to look at nuclear DNA, and
this is important for a couple of reasons. It allows us to genetically
sex individuals; this is important when you have very fragmentary remains,
or remains of young individuals who can't be sexed just by looking at the
anatomy. And this allows us to get at, again, social structure, status,
there's forensic science applications, and also patterns, differential
patterns by gender of disease, diet, material possessions, and that sort
And then, of course, there's the question that a lot of us here are
concerned with, and that is identifying prehistoric groups and their relationships
to modern groups. With large sample sizes, the frequencies of these
haplogroups and D-loop mutations can be used to characterize populations.
Because populations inherit these markers from their ancestors, although
social boundaries are permeable to genes, they're not completely permeable.
So you do have different frequencies of these markers in different groups.
And so groups that are closely related are likely to have very similar
frequencies, and groups that don't have, groups that have very, very different
frequencies, are not likely to be closely related. An example of
how we can use this -- oops, I'm missing a slide. Well, okay, here
you go, mapping the past. An example of how we can use this is in
the Pacific, there's a marker in the mitochondrial DNA called the nine
base pair deletion, and its frequency is, obviously, it varies quite a
bit. It's very frequent out in the Polynesian islands, very infrequent
in Australia, New Guinea and some other groups, and by examining, looking
for the presence or absence of this marker in ancient individuals in these
islands, this can help us understand how these islands were peopled, from
where they were peopled, how they interacted with each other after they
were peopled, how these genetic groups correlate with linguistic groups,
that sort of thing.
So those are the types of questions I'm interested in. They may
not be the types of questions you're interested in, and that's part of
a later issue we're going to discuss. That was the promise of ancient
DNA. Now, for the problems! There are some obvious methodological
problems working with ancient DNA, and the major one is contamination.
Because of the technology that allows us to access this ancient DNA, we
also have to worry about getting contamination from modern sources.
This may be the archaeologist who excavated the remains, the museum curator
who moved them from one drawer to another, the technician who packaged
the tubes in which you're storing the DNA, the person in the laboratory
who sneezed on his or her gloves and then handled them, other ancient individuals
who you have amplified and then spilled on the lab bench, their DNA.
So it's essential when you're working with ancient DNA to have very stringent
contamination controls, to decontaminate the surfaces of the samples before
you look at them, and of course to monitor for any evidence of contamination.
Then sample size, as I mentioned earlier, is a huge issue here, especially
when you're dealing with defining groups. To understand prehistoric
behavior, to get an accurate picture of the ancestor-descendant relationships,
requires that we look at the range of behavior and relationships in a group,
and get an accurate estimate of these frequencies. You cannot do
this unless you have a large sample size. A sample size of one doesn't
get you a frequency of anything. A frequency of 100% in a sample
size of one doesn't tell you what else was present in that group of people
at the time that this individual lived. So sample size from ancient
groups is often limited, because you don't find very many ancient individuals.
In addition, sample size from modern groups is an issue, because you can't
connect an ancient group with a modern group unless you can compare them
to that data in a modern group. And this slide depicts -- is backwards!
-- and depicts that frequency of that nine base pair deletion. It's
okay, I don't need it forwards, I can tell you what it says. The
frequency of that deletion that I talked about in Polynesia? Well,
it's present across Asia and the Americas, and we've got the frequency
in South American groups, Central, North, Siberian, Southeast Asian, Oceanic
groups. And as you can see, it does vary quite a bit from group to
group. But if we could actually read the names down there, of these
groups, you would notice a whole heck of a lot of groups aren't there.
We don't have any idea what the frequencies in these groups are.
The other thing, if you could read this, you would notice, is, for example,
you've got some groups that don't seem to have the deletion at all.
But if you could read the sample sizes, they would be 10, 12 individuals
from a whole population. Well, if the frequency of that deletion
is 10% 6%, only sampling 10 individuals from that population is likely
not to catch that. So, again, sample size is a big issue.
And then data analysis is another big issue. There's a long history
of looking at modern genetic variation and there are many methods developed
to do this. But it's unclear how they can be profitably and successfully
applied to samples that span not only space, that are from different geographic
areas, but as sampled across time, because ancient groups, when you sample
them, often you get one individual from each generation, or one individual
perhaps every hundred years. So we're only just beginning the empiric
-- how we can do this using empirical data, simulation data, and extensions
of modern theory.
So that's usually where a review of ancient DNA in the literature would
stop. But obviously, we haven't gone far enough, because there's
a qualitatively different set of questions. Because when we deal
with human ancestors, there are living peoples with widely different social,
political, religious and legal beliefs and interests. Although we
may be advancing the pursuit of knowledge (or may not be), we are certainly
going to be affecting the lives of the modern members of our species when
we study these ancient individuals. So our results, especially in
today's climate of genetic essentialism, are often going to be interpreted
to have serious social meaning and policy implications, and we need to
know what those are before we decide to undertake a study. An obvious
example of this is the Kennewick Man remains, which were discovered a few
years ago off the Columbia, on the banks of the Columbia River, and they
are the remains of about a 9,000 year-old individual. And today's
contemporary attitudes towards the importance of genetic study, and all
the media attention lately on "really cool" ancient DNA stuff, like Neanderthal
DNA, has seemed to lead to a bias toward science in the covering of this
story. It's even been covered in science fiction magazines.
There was an article in this October/November 1998 Asimov Science Fiction,
where we got a "Science Fact" article about Kennewick. And in your
reader, there was a New Yorker article written about the Kennewick case.
A bunch of scientists have filed suit to protect the remains against repatriation,
on the insistence that it must be studied, whereas the federal government
insists that under NAGPRA law, it should be repatriated and reburied, or
not, depending on what Native Americans decide. In this 11-page article,
the author, Douglas Preston, directly quoted 13 scientists: anthropologists,
archaeologists, geneticists, geologists; and he directly quoted, or even
mentioned, one Native American, and didn't seem to have interviewed anybody
at the Army Corps of Engineers, the federal agency responsible for this
So, obviously, we're dealing with a situation where the attitudes towards
ancient study differ from different groups, and the Western attitude favoring
scientific study seems to be the one that's the popular attitude in popular
science, popular literature and media. So we know now, or we're starting
to know now, that we can't just take this attitude. Science is not
performed in a vacuum, and so these attitudes have begun to be challenged.
Because they have implications -- discrimination, racism -- our results
may contradict deeply held beliefs, and of course may affect legal claims.
Now, this is true of modern genetic variation studies too. Ancient
DNA has a separate or related set of problems, because you can't consult
the ancient individual, obviously, which means you're left really with
two choices: consult no one, which, we seem to be arriving at the conclusion,
is not generally an acceptable attitude; or consult the modern descendants
or some modern group who has an interest in this ancient individual.
But that's easier said than done. Usually, or in many cases, we're
not sure who that modern descendant is. In fact, in a lot of cases,
the study is undertaken to help find that out.
So, what evidence do we use to decide who to ask? And if that
evidence is contradictory, what do we go with? With Native Americans,
that particular issue seems to be settled with NAGPRA, except of course
for various interpretations of NAGPRA. But it's obviously not settled
with most other groups. What if, when you're doing your study, your
preliminary evidence suggests that you've asked the wrong group?
For example, in Kennewick, the Kennewick Man case, had the Native Americans
given permission for study, and the genetic evidence suggested that this
particular individual didn't fall into the general variation found among
Native Americans, and instead fell within the variation found within some
other group? Should we stop that study right then and go to that
other group and ask them, "Should we continue?" What if they say
no? Should we stop? If they say yes, and it turns out we were
wrong with our preliminary conclusions, what do we do?
These are all questions that are of interest to those of us who study
ancient DNA, and also those of us who are descended from these individuals,
and I'm hoping that we're going to get somewhere on this today, and Dennis
is going to talk a little bit about his efforts to figure this sort of
thing out. Thanks.
*This talk has been edited for web publishing by the author.