Re: Human/chimp similarity measures

Stephen E. Jones (
Thu, 07 Oct 1999 20:51:20 +0800


On Wed, 06 Oct 1999 10:34:33 -0700, Cliff Lundberg wrote:


>HX>The first three techniques are based upon small-scale
>>sampling, which Stephen cites as a difficulty. DNA
>>hybridization, though, is based upon large-scale sampling,
>>involving much longer stretches of DNA than are referenced by
>>any of the other techniques. DNA hybridization provides a
>>comparison without reference to whether the sequence being
>>analyzed is composed of coding regions, non-coding regions,
>>or a mixture of the two.

CL>Can someone provide more detail on how this works? It seems to me that
>if one doesn't use *all* the DNA in the nuclei, that comparisons will be
>inevitably skewed by the sampling of different regions. And the measurement
>of heat required seems too crude a technique to resolve the matters in

It seems to me to be a "crude technique", as Gribbin and Cherfas describe it
below. Any large strands of the two DNAs which don't match are just discarded
as anomalies ("Once the mixture is cooled the few remaining single strands
are removed")! It *presupposes* a close fit and then finds, incredibly :-),
that there *is* a close fit!:

"The idea behind DNA hybridization is simplicity itself. We separate the
double helix of one species' DNA into its component single complementary
strands. We do the same to the DNA of the species we wish to compare it
with. Then we mix all the separated strands together. Where the two
species have identical sequences along their DNA, the complementary
bases from the two different species will be able to come together and join.
In fact, if the two sets of DNA were completely identical, then we would
end up with three different sorts of DNA after the mixing. Two sorts would
be identical to the original, composed of complementary strands that had
initially been in the same sample of DNA, though we would be very
surprised if any particular strand managed to join with its original partner.
Some might, of course, but in any case this doesn't matter as the double
strands would be identical in every respect with the original sample, even if
the pairs that come together in each new double helix are not original pairs.
In addition to this reconstituted DNA we would also find in our sample an
equal amount of DNA in which one strand had come from one sample and
the other from the other, but unless we had previously labelled one of the
lots of DNA we would not be able to tell which double-stranded molecule
was which. If one sample had been made highly radioactive, and if we
could count the radioactivity of each newly-formed double helix we would
find that a quarter of the molecules were as radioactive as the original
sample, and another quarter were not at all radioactive. These would be the
molecules that had by lucky chance joined up with a complementary strand
from their own sample. The remaining half would have half the
radioactivity of the original sample, and these would be double-stranded
molecules composed of one strand from the labelled sample and the
complementary strand from the unlabelled sample.

That is how things would work if the two samples of DNA were in fact
absolutely identical. What would happen if this were not the case? Well, if
the two samples were very similar, so that long stretches of each did indeed
contain almost the same sequence of bases, then they would still be able to
form so-called hybrid molecules . What we then want to know is the extent
of the similarity, the percentage of the two DNAs that is in fact identical.
To do this we make use of the fact that it is only the binding between
individual pairs of complementary bases that holds the two strands

The more bonds there are, the harder it will be to separate the strands of
the hybrid DNA, and to measure the strength of the bonding between the
strands we need only put in energy to overcome the binding energy and see
when the strands drift apart. What we are really doing is finding the melting
point of the DNA. Solids are held together by bonds between their
component atoms. Inject energy into the solid, in the form of heat, and you
break the bonds so that the solid is free to become a liquid. The energy
needed to break the molecular bonds is the melting point of the substance.
To get back to DNA, if the two strands are totally complementary they will
be held together with a certain strength and it will take a certain amount of
energy, a temperature of around 85 degrees Celsius, to separate them. If
the strands are only partially complementary the force holding them
together will be weaker and it will take less energy, a lower temperature, to
make them drift apart. Impure DNA, in the sense of being made of non-
identical strands, has a lower melting point than pure DNA, just as impure
water has a lower melting point (freezes at a lower temperature) than pure

Naturally, like all the techniques we have discussed, the theory of DNA
hybridisation is easy enough to understand but putting it into practice is the
devil's own job. Quite apart from the problems of extracting and purifying
the DNA of the species we want to compare, there is the very tricky
business of detecting the melting point of the hybrid. It isn't as simple as
keeping one eye on a thermometer and the other on a block of ice, for
although DNA can actually be seen under an electron microscope after
suitable preparation it is impossible to tell when the strands in a solution of
DNA have separated. Once the scientist has made his hybrid molecules,
also known as heteroduplex DNA (which just means double-stranded DNA
from different sources) he has to measure very accurately the temperature
at which the double helices break down. The procedures depend heavily on
the enormous advances that have been made in the technology of genetic
engineering, a full discussion of which would be out of place here. We can,
nevertheless, give some idea of how it is done.

The first step, as we've said, is to purify the DNA from the two species.
Then, one species' DNA is labelled with a radioactive tag, usually the
radioactive isotope of iodine; this does not affect the way the strands work,
but simpler acts as a tracer that enables researchers to follow the single
strands from that species. With the labelling done, the two sets of DNA are
mixed and slowly heated. At around 85 degrees Celsius the bonds between
opposite bases, which normally hold the strands together, are broken, and
the strands drift apart. Now the mixture is allowed to cool slowly so that
heteroduplex molecules can form from the two species of DNA. Once the
mixture is cooled the few remaining single strands are removed and the
business of measuring the melting point begins The temperature is raised by
about one degree and the dissociates DNA removed and assayed for
radioactivity with an improves version of the old-fashioned geiger counter.
Then the temperature is raised another notch and the next lot of single
strands removed ant counted. A repeated series of counts at steadily
increasing tempera turns produces a so-called dissociation curve, the peak
of which represents the melting point of the hybrid DNA. This can then be
compared directly with the melting point of a pure hybrid, that is DNA
from the target species heated and allowed to recombine, so that any quirks
due to the heating processes and so on are evened out. The size of the
difference in melting points between heteroduplex and normal DNA is
directly related to the dissimilarity between the two strands. A difference of
one degree Celsius is roughly equivalent to a difference in one per cent of
the DNA; one in a hundred of the nucleotides are not identical in two
species that show a melting point depression of one degree Celsius."

(Gribbin J. & Cherfas J., "The Monkey Puzzle: A Family Tree"
1982, pp97-100)



"The code of conduct that the naturalist wishing to understand the problem
of evolution must adopt is to adhere to facts and sweep away all a priori
ideas and dogmas. Facts must come first and theories must follow. The
only verdict that matters is the one pronounced by the court as proved
facts. Indeed, the best studies on evolution have been carried out by
biologists who are not blinded by doctrines and who observe facts coldly
without considering whether they agree or disagree with their theories."
(Grasse P.-P., "Evolution of Living Organisms: Evidence for a New
Theory of Transformation", Academic Press: New York NY, 1977, p8)