4.5.5 Comparison of DNA Sequences from the Thermostability of Heteroduplices
153
given genes from groups X and Y, respectively,
are identical; Jxy refers then to the comparison of
a gene from group X with one from group Y. I can
have values between 0 and 1; I = 1 if identical
alleles occur with the same frequency at all loci of
X and Y, and I = 0 if this occurs at no locus [292].
The estimation of genetic distance using electrophoretic data is not free of problems:
1. D depends upon the choice of the loci to be
compared, and increases as the heterozygosity
of the chosen loci increases; thus, highly polymorphic loci also show large interspecific differences [62, 381].
2. Even otherwise insignificant gene exchange
between two populations X and Y hinders the
development of a large genetic distance.
Assuming that per generation mx individuals
migrate from X to Y and my migrate in the
opposite direction, then, according to the
neutral theory, the genetic identity is given by
I = (mx + my)/(mx + my + v), (4.19)
where v is the number of mutations arising per
generation at the relevant locus [292].
The measurement of distance DR of Roger (1972)
is also relatively frequently applied; in contrast to
Nei's D value, DR conforms to the so-called triangle condition that is important for the construction of phylogenetic trees, i.e. the estimated three
distance values between three animal groups
form a triangle:
DR = (lIr) Lj [Y2Li (Xij - Yij)2]"'2. (4.20)
DR can have values between 0 and 1 and is usually
somewhat smaller than D [292].
The evolutionary distance values obtained
from electrophoretic data agree fairly well with
distances determined immunologically or by
DNA hybridization, but are inevitably subject to
error because different loci usually have different
rates of substitution [175]. D and I values
obtained from I-D electrophoresis are mainly
suitable for intraspecific comparisons or those
between closely related species; 2-D electrophoresis is more suitable for the comparison of distantly related taxa [391].
4.5.5 Comparison of DNA Sequences from
the Thermostability of Heteroduplices
DNA double helices dissociate on heating into
single strands which reassociate on cooling; the
value of the dissociation or melting temperature
(Tm) depends upon the strength of the interaction
of the complementary strands. A heteroduplex
made up of DNA single strands from different
species has a lower Tm than fully complementary
DNA; the difference dTm is a measure of the
sequence difference. Until recently, it was
accepted that a d T m of 1 °C corresponded to a
sequence difference of about 1 % ; however, new
measurements suggest 1.7 % [53]. Reassociation
of total eukaryotic DNA only really involves
repetitive DNA which, because of its limited
information content and rapid rate of evolution,
is unsuitable for the analysis of relationships.
Therefore, since 1970, fragmented single-copy
DNA has been used for such measurements.
The estimation of DNA sequence differences
by these methods is associated with several difficulties and sources of error [44]:
1. An error arises in d T m determination because
the latter is dependent upon the base composition; however, this effect can be mitigated by
careful choice of the buffer.
2. Tm is dependent upon the length of the DNA
fragments; this can be corrected for mathematically.
3. Distinct DNA polymorphism, e.g. in the echinoderms, leads to the situation whereby reassociated DNA from two or more individuals
already shows a 2-4 °C lower T m compared
with duplex DNA from one individual [151].
4. Reassociation is never complete; this can also
be corrected for mathematically. However, in
some Drosophila species, one-third of the
DNA takes no part in duplex formation,
whereas the remaining two-thirds shows great
sequence similarity; in some starfish, only
one-third of the DNA reacts at all. In such
cases, and generally for high dTm values, the
accuracy of this form of sequence comparison
should not be overestimated.
The genetic distances derived from DNA hybridization are usually significantly larger than those
obtained by other methods. Thus, in the case of
four Drosophila species from three neighbouring
Hawaiian islands, enzyme-electrophoresis and
d T m determinations gave the same picture of
relationships during colonization of the islands
(Fig. 4.11), but the electrophoretically determined genetic distances were significantly
smaller, perhaps as the result of selection at the
investigated loci [186]. Amongst the various
methods for the analysis of relationships by
investigations of the DNA, the sequencing of
individual genes stands at one extreme and Tm
153
given genes from groups X and Y, respectively,
are identical; Jxy refers then to the comparison of
a gene from group X with one from group Y. I can
have values between 0 and 1; I = 1 if identical
alleles occur with the same frequency at all loci of
X and Y, and I = 0 if this occurs at no locus [292].
The estimation of genetic distance using electrophoretic data is not free of problems:
1. D depends upon the choice of the loci to be
compared, and increases as the heterozygosity
of the chosen loci increases; thus, highly polymorphic loci also show large interspecific differences [62, 381].
2. Even otherwise insignificant gene exchange
between two populations X and Y hinders the
development of a large genetic distance.
Assuming that per generation mx individuals
migrate from X to Y and my migrate in the
opposite direction, then, according to the
neutral theory, the genetic identity is given by
I = (mx + my)/(mx + my + v), (4.19)
where v is the number of mutations arising per
generation at the relevant locus [292].
The measurement of distance DR of Roger (1972)
is also relatively frequently applied; in contrast to
Nei's D value, DR conforms to the so-called triangle condition that is important for the construction of phylogenetic trees, i.e. the estimated three
distance values between three animal groups
form a triangle:
DR = (lIr) Lj [Y2Li (Xij - Yij)2]"'2. (4.20)
DR can have values between 0 and 1 and is usually
somewhat smaller than D [292].
The evolutionary distance values obtained
from electrophoretic data agree fairly well with
distances determined immunologically or by
DNA hybridization, but are inevitably subject to
error because different loci usually have different
rates of substitution [175]. D and I values
obtained from I-D electrophoresis are mainly
suitable for intraspecific comparisons or those
between closely related species; 2-D electrophoresis is more suitable for the comparison of distantly related taxa [391].
4.5.5 Comparison of DNA Sequences from
the Thermostability of Heteroduplices
DNA double helices dissociate on heating into
single strands which reassociate on cooling; the
value of the dissociation or melting temperature
(Tm) depends upon the strength of the interaction
of the complementary strands. A heteroduplex
made up of DNA single strands from different
species has a lower Tm than fully complementary
DNA; the difference dTm is a measure of the
sequence difference. Until recently, it was
accepted that a d T m of 1 °C corresponded to a
sequence difference of about 1 % ; however, new
measurements suggest 1.7 % [53]. Reassociation
of total eukaryotic DNA only really involves
repetitive DNA which, because of its limited
information content and rapid rate of evolution,
is unsuitable for the analysis of relationships.
Therefore, since 1970, fragmented single-copy
DNA has been used for such measurements.
The estimation of DNA sequence differences
by these methods is associated with several difficulties and sources of error [44]:
1. An error arises in d T m determination because
the latter is dependent upon the base composition; however, this effect can be mitigated by
careful choice of the buffer.
2. Tm is dependent upon the length of the DNA
fragments; this can be corrected for mathematically.
3. Distinct DNA polymorphism, e.g. in the echinoderms, leads to the situation whereby reassociated DNA from two or more individuals
already shows a 2-4 °C lower T m compared
with duplex DNA from one individual [151].
4. Reassociation is never complete; this can also
be corrected for mathematically. However, in
some Drosophila species, one-third of the
DNA takes no part in duplex formation,
whereas the remaining two-thirds shows great
sequence similarity; in some starfish, only
one-third of the DNA reacts at all. In such
cases, and generally for high dTm values, the
accuracy of this form of sequence comparison
should not be overestimated.
The genetic distances derived from DNA hybridization are usually significantly larger than those
obtained by other methods. Thus, in the case of
four Drosophila species from three neighbouring
Hawaiian islands, enzyme-electrophoresis and
d T m determinations gave the same picture of
relationships during colonization of the islands
(Fig. 4.11), but the electrophoretically determined genetic distances were significantly
smaller, perhaps as the result of selection at the
investigated loci [186]. Amongst the various
methods for the analysis of relationships by
investigations of the DNA, the sequencing of
individual genes stands at one extreme and Tm