Biology Reference
In-Depth Information
neighbors. The network of open circles stands for a large
connected genotype network that traverses genotype space.
The colored circles represent genotypes whose phenotype
is different from P (one color per phenotype), and that are
neighbors of the genotypes on the genotype network. Note
that different regions of this hypothetical genotype space
contain different colors. The two large dashed circles
represent the neighborhoods of two genotypes on the
genotype network. Note that the phenotypes (colors) in
these two neighborhoods are different, a reflection of
property 2. Note that
Figure 13.2
represents a complex,
vast, and high dimensional genotype space in a highly
simplified, two-dimensional way. For example, actual
genotype networks contain an astronomical number of
members. Individual genotypes may have hundreds to
thousands of neighbors, only few of which can be shown. In
addition, each of the colored genotypes is also part of a vast
genotype network that is not shown.
networks and their neighborhoods allow populations to
preserve old phenotypes while exploring many new
phenotypes.
Neither property 1 nor property 2 alone would be
sufficient for such exploration
[10]
. Without property 1 (no
genotype networks), a population would have low geno-
typic diversity and could therefore not explore different
neighborhoods in this space. The total number of pheno-
types is much greater than the number of phenotypes in
a neighborhood for any one of the three system classes
[10]
.
Thus, the absence of genotype networks would mean that
most novel phenotypes are off-limits to an evolving pop-
ulation. Conversely, in the absence of property 2, that is, if
the neighborhoods of different genotypes contained mostly
identical new phenotypes, the existence of genotype
networks would be irrelevant to the exploration of novel
phenotypes. The reason is that even though a population's
genotypes could change during evolutionary exploration of
a genotype network, the changing genotypes would have
access to the same unchanging spectrum of novel
phenotypes.
The second question posed earlier regards the multiple
evolutionary origins of many evolutionary innovations
[6,
53]
. Such multiple origins may be difficult to understand, if
one assumes that innovations are unique solutions to
particular problems that life faces, and that they are unique
because the underlying problems are difficult to solve.
Viewing such solutions from the vantage point of a geno-
type space leads to a completely different perspective.
There, a genotype with a specific phenotype can be viewed
as a solution to a particular problem. The existence of vast
genotype networks for typical phenotypes means that
typical problems have not just one, but astronomically
many solutions. Different genotypes on the same genotype
network can be viewed as different solutions to the same
problem. Populations of organisms that explore genotype
space from different starting points may encounter different
solutions. To be sure, most innovations may involve
multiple changes in all three major system classes, but
because genotype networks are ubiquitous in all three
classes, so are multiple solutions to most problems. From
this perspective, the multiple origins of many evolutionary
innovations are not surprising but rather to be expected.
The third question is whether innovation is usually
combinatorial, involving old parts that are combined to new
purposes. Here again, the vantage point of a genotype
space, which contains all possible innovations, suggests
a very straightforward answer: all innovation is combina-
torial. New functions of proteins emerge through new
combinations of amino acids. New metabolic phenotypes
emerge through new combinations of already existing
biochemical reactions. And new gene expression patterns
of regulatory circuits arise through new combinations of
regulators and their interactions.
Genotype Networks and Their Diverse
Neighborhoods Can Help Explain the Origin
of New Phenotypes
I will now return to the three questions about the origin of
new phenotypes posed earlier, and which a systematic
understanding of phenotypic variability needs to address.
The first is that organisms need to preserve old, well-
adapted phenotypes while exploring many new phenotypes.
Properties 1 and 2 can jointly help answer this question. To
see this, consider that all evolution takes place in pop-
ulations of organisms, each with its own genotype. Envi-
sion a population of genotypes in any one of our three
system classes. Individuals in this population have
a phenotype that may be necessary for their survival, but
somewhere in genotype space a superior phenotype may
exist. The genotypes of individuals in this population suffer
mutations that affect their genotype. Natural selection
eliminates any mutants that have not preserved the old
phenotype or replaced it with a superior phenotype. One
can view such a population as a cloud of points
[112]
that
diffuses on a genotype network through genotype space.
Genotype networks (property 1) allow the genotypes of
individuals in such a population to change without affecting
their phenotype. They allow the preservation of old
phenotypes despite genotypic change. Over time, geno-
types may change dramatically while preserving their
phenotype. During this process, the population explores
different regions of genotype space. Because of property 2,
the diversity of genotypic neighborhoods, the neighbor-
hoods of the population's genotypes will contain ever-
changing sets of new phenotypes. This means that the
population can explore different novel phenotypes in its
neighborhood as its genotypes change. In sum, genotype
