4.5.7 Construction of Phylogenetic Trees from Molecular Data
155
in the DNA. From the wealth of discussion and
large number of suggestions for solutions in the
literature [29], it is clear that the numerical
accuracy of genetic distance values obtained by
mtDNA restriction analysis should not be overestimated; with very large genetic distances
(6 > 0040) the method is completely unusable.
The use of mtDNA in the analysis of relationships
presents some special problems. Thus, the considerable polymorphism of mtDNA has a disturbing effect. By the chance crossing-in of a foreign
female, the mtDNA can be replaced by that of
another species, and because mtDNA is passed
on maternally this situation then persists in the
population (pp.166 and 168).
DNA relationship analysis with the potential of
extremely high resolution has now become possible through the introduction of DNA fingerprinting. This method uses as a probe a certain
type of repetitive DNA that is widely dispersed in
the genome, the so-called mini-satellite. Total
DNA of suitable cells (e.g. leukocytes) is cleaved
by restriction enzymes into defined fragments,
the fragments are separated electrophoretically
and hybridized on a Southern blot with a radioactively labelled mini-satellite. The resulting pattern is specific for each individual; the similarity
increases with increasing relationship, being
highest between parents, children and siblings. A
greater degree of separation, as is required, for
example, for parenthood analysis or forensic purposes, can be obtained by 2-D electrophoresis
[418]. Highly variable mini-satellites are apparently widespread and the method is, therefore,
increasingly used for population genetics
investigations on vertebrates of all classes
[155, 255, 343, 416].
4.5.7 Construction of Phylogenetic Trees
from Molecular Data
The evolutionary relationship between homologous genes or proteins can be represented by a
phylogenetic tree. This can be looked upon as a
genealogical tree of the molecules but also of the
species from which the molecules were obtained.
A molecular phylogenetic tree describes which
nucleotide substitutions or amino acid exchanges
have occurred in the course of evolution. This, on
the one hand, provides data from which the laws
of molecular evolution may be derived and, on
the other hand, can be viewed as a natural experiment that allows conclusions to be drawn about
the functions of the individual components of the
macromolecules. A species phylogenetic tree presents special possibilities because macromolecules can be compared more widely than morphological characters, in fact, often right across the
whole animal kingdom or across all organisms.
A particular form of molecular relationship arises through gene duplication, which is a common
event in molecular evolution. The genes resulting
from duplication are, of course, homologous but
may become quite different during the course of
further evolution. It is therefore necessary in
comparing homologous genes or gene products to
distinguish between paralogous genes, which
belong to two different evolutionary lines arising
by gene duplication, and orthologous genes,
which are of the same evolutionary line. Whereas
a molecular phylogenetic tree may include both
paralogous and orthologous molecules, only
orthologous molecules may be used in the construction of a species phylogenetic tree. A confusing situation may arise when only one of two gene
lines, arising by duplication, is retained in a particular species. The best-known example of this is
the lysozyme of avian eggs; this is very different
in the goose and the chicken despite the close
relationship of these species. This puzzle was only
solved when the enzymes of the goose and the
chicken type were discovered together in the
black swan, Cygnus atratus, presenting evidence
for a duplication of the lysozyme gene (p. 504);
thus, the lysozymes from eggs of the goose and
chicken are paralogous.
Qualitative variation in individual characters
may be used for the construction of molecular
phylogenetic trees, as in classical phylogenetics,
or quantitative data, as in numerical taxonomy.
Numerical methods are preferred in studies of
molecular relationships for two reasons: (1) only
phylogenetic trees constructed with the help of
quantitative methods can be tested statistically;
and (2) according to prevalent ideas, the influence of selection can be largely neglected in phylogenetic tree construction using molecular characters, and the rate of evolution may be taken to
be constant over long periods (molecular clock
hypothesis). Thus, the expression: "the more
similar, the more closely related", which is easily
refutable for morphological characters, is actually
more appropriate for the molecular analysis of
relationships. Many of the molecular data are
quantitative from the outset, e.g. immunological
distance, amino acid composition or ilTm values.
Whenever essentially qualitative data, e.g. from
sequencing, restriction analysis or electrophoresis, are converted to a quantitative measure of
155
in the DNA. From the wealth of discussion and
large number of suggestions for solutions in the
literature [29], it is clear that the numerical
accuracy of genetic distance values obtained by
mtDNA restriction analysis should not be overestimated; with very large genetic distances
(6 > 0040) the method is completely unusable.
The use of mtDNA in the analysis of relationships
presents some special problems. Thus, the considerable polymorphism of mtDNA has a disturbing effect. By the chance crossing-in of a foreign
female, the mtDNA can be replaced by that of
another species, and because mtDNA is passed
on maternally this situation then persists in the
population (pp.166 and 168).
DNA relationship analysis with the potential of
extremely high resolution has now become possible through the introduction of DNA fingerprinting. This method uses as a probe a certain
type of repetitive DNA that is widely dispersed in
the genome, the so-called mini-satellite. Total
DNA of suitable cells (e.g. leukocytes) is cleaved
by restriction enzymes into defined fragments,
the fragments are separated electrophoretically
and hybridized on a Southern blot with a radioactively labelled mini-satellite. The resulting pattern is specific for each individual; the similarity
increases with increasing relationship, being
highest between parents, children and siblings. A
greater degree of separation, as is required, for
example, for parenthood analysis or forensic purposes, can be obtained by 2-D electrophoresis
[418]. Highly variable mini-satellites are apparently widespread and the method is, therefore,
increasingly used for population genetics
investigations on vertebrates of all classes
[155, 255, 343, 416].
4.5.7 Construction of Phylogenetic Trees
from Molecular Data
The evolutionary relationship between homologous genes or proteins can be represented by a
phylogenetic tree. This can be looked upon as a
genealogical tree of the molecules but also of the
species from which the molecules were obtained.
A molecular phylogenetic tree describes which
nucleotide substitutions or amino acid exchanges
have occurred in the course of evolution. This, on
the one hand, provides data from which the laws
of molecular evolution may be derived and, on
the other hand, can be viewed as a natural experiment that allows conclusions to be drawn about
the functions of the individual components of the
macromolecules. A species phylogenetic tree presents special possibilities because macromolecules can be compared more widely than morphological characters, in fact, often right across the
whole animal kingdom or across all organisms.
A particular form of molecular relationship arises through gene duplication, which is a common
event in molecular evolution. The genes resulting
from duplication are, of course, homologous but
may become quite different during the course of
further evolution. It is therefore necessary in
comparing homologous genes or gene products to
distinguish between paralogous genes, which
belong to two different evolutionary lines arising
by gene duplication, and orthologous genes,
which are of the same evolutionary line. Whereas
a molecular phylogenetic tree may include both
paralogous and orthologous molecules, only
orthologous molecules may be used in the construction of a species phylogenetic tree. A confusing situation may arise when only one of two gene
lines, arising by duplication, is retained in a particular species. The best-known example of this is
the lysozyme of avian eggs; this is very different
in the goose and the chicken despite the close
relationship of these species. This puzzle was only
solved when the enzymes of the goose and the
chicken type were discovered together in the
black swan, Cygnus atratus, presenting evidence
for a duplication of the lysozyme gene (p. 504);
thus, the lysozymes from eggs of the goose and
chicken are paralogous.
Qualitative variation in individual characters
may be used for the construction of molecular
phylogenetic trees, as in classical phylogenetics,
or quantitative data, as in numerical taxonomy.
Numerical methods are preferred in studies of
molecular relationships for two reasons: (1) only
phylogenetic trees constructed with the help of
quantitative methods can be tested statistically;
and (2) according to prevalent ideas, the influence of selection can be largely neglected in phylogenetic tree construction using molecular characters, and the rate of evolution may be taken to
be constant over long periods (molecular clock
hypothesis). Thus, the expression: "the more
similar, the more closely related", which is easily
refutable for morphological characters, is actually
more appropriate for the molecular analysis of
relationships. Many of the molecular data are
quantitative from the outset, e.g. immunological
distance, amino acid composition or ilTm values.
Whenever essentially qualitative data, e.g. from
sequencing, restriction analysis or electrophoresis, are converted to a quantitative measure of