Transcript Hybrid

Lecture VI Molecular Ecology
(Hybridization)
What is Hybridization?
Hybrid – Oxford Dictionary of Zoology –
- an individual animal that results from a cross between parents of
differing genotypes. Strictly, most individuals in an outbreeding
population are hybrids, but the term is more usually reserved for cases
in which the parents are individuals whose genomes are sufficiently
distinct for them to be recognized as different species or subspecies.
A good example is a mule, produced by cross-breeding an ass and a
horse (each of which can breed as true species). Hybrids may be fertile
or sterile depending on qualitative and/or quantitative differences in the
genomes of the two parents. Hybrids, like the mule, whose parents are
of different species, are frequently sterile.
How important is hybridization?
Why devote a whole lecture to it?
The theoretical and empirical study of hybridization is somewhat
inseparable from that of speciation. Remember the BSC (Biological
Species Concept). If reproductive islolation should define a species,
then what does the existence of hybridization mean? Does it mean that
two lineages that hybridize are not species?
Hybridization is one of the primary problems in applying the BSC, the
other being that knowledge on reproductive isolation (or, inability to
hybridize) is often not available.
For fishes, this issue is even more prominent as not only do many
fishes hybridize naturally in nature, but those that don‘t will often do so
under artificial conditions, and even more often if given a „little help“
in the laboratory.
Hybrid – Oxford Dictionary of Zoology –
- an individual animal that results from a cross between parents of
differing genotypes. Strictly, most individuals in an outbreeding
population are hybrids, but the term is more usually reserved for cases
in which the parents are individuals whose genomes are sufficiently
distinct for them to be recognized as different species or subspecies.
A good example is a mule, produced by cross-breeding an ass and a
horse (each of which can breed as true species). Hybrids may be fertile
or sterile depending on qualitative and/or quantitative differences in the
genomes of the two parents. Hybrids, like the mule, whose parents are
of different species, are frequently sterile.
In practice, evolutionary biologists do not draw any „systematic lines“,
when discussing hybridization. As you have learned, it is difficult enough
to define species, with a „subspecies being even more controversial. Thus,
in practice, the taxonomic level of the units described to hybridize is very loosely
interpreted, and with domesticated lines, for example, one may related simply to
a single trait, or genetic locus.
Introgression – (the distinction between hybridization &
introgression is very important!!)
The incorporation of the genes of one species into the gene pool
of another species. If the ranges of two species overlap and
fertile hybrids are produced, the hybrids tend to backcross with
the more abundant species. This results in a population in
which most individuals resemble the more abundant parents but
also possess some of the characteristics of the other parent
species.
Hybrid vigour (heterosis) – (primarily relevant in domesticated
strains, and is probably mostly do to a „release“ from the effects
of inbreeding)
The increased vigour of growth, survival, and fertility of hybrids,
as compared with the two homozygotes. It usually results from
crosses between two genetically different, highly inbred lines. It
is always associated wtih increased heterozygosity.
Hybrids swarm –
A continuous series of hyrids that are morphologically distinct from
one another, which results from the hybridization of two species
followed by the crossing and backcrossing of subsequent generations.
The hybrids are very variable owing to segregation of alleles at each
locus.
Hybrid dysgenesis–
A complex of genetic abnormalities, which occurs in certain hybrids.
The abnormalities may include sterility, enhanced rates of gene
mutations, and chromosomal rearrangements. Hybrid dysgenesis occurs
in the hybrid offspring of certain strains of Drosophila, in which it is
thought to be due to mutations induced by transposon-like elements.
How Common is Hybridization in Fishes?
Schwartz (1972, 1981) compiled a total of 3,759 references dealing with the natural and artificial
hybridization of fishes.
Schwartz FJ 1981. World Literature to Fish Hybrids, With an Analysis by Family, Species, and
Hybrid: Supplement 1. NOAA Technical Report NMFS SSFR-750, US Dept. of Commerce 507 pp.
Most of these studies were published after Hubbs (1955) summarized his extensive investigations of
natural hybridization among North American fishes.
Hubbs CL (1955) Hybridization between fish species in nature. Systematic Zoology 4, 1-20.
Natural hybridization is believed to be more common in fishes than in other groups of
vertebrates.
Several characteristics of fishes may account for this distinction:
1) external fertilization
2) weak ethological isolating mechanisms,
3) unequal abundance of the two parental species,
4) competition for limited spawning habitat,
5) susceptibility to secondary contact between recently evolved forms
These characteristics are affected to varying degrees by local habitat.
I would add artificial rearing, transport and planned or accidental releases to this list. Of these
characteristics, which are promoting by environmental changes, natural or man-made?
Natural and man-induced changes in environmental conditions are often cited as causes of
hybridization in fishes. For example, hybridization is relatively common among temperate
freshwater fishes in areas where geologic and climatic across the Pleistocene have
drastically altered aquatic environments, but hybridization appears to be rare among
marine and tropical fishes that inhabit more stable environments. Man caused habitat
changes in North America have also been correlated with hybridization between both
previously allopatric and naturally sympatric pairs of species (Hubbs et al 1953, Nelson
1966, 1973, Stevenson & Buchanon 1973). In addition, sympatric species that rarely or
never hybridize in nature often hybridize freely within the confines of aquaria (Hubbs
1955). As Hubbs's (1955) extensive investigations led him to conclude " It is evident that the
hybridization is conditioned by environmental factors"-
Thus, it is important to understand the role of the environment in
hybridization.
The problems with hybrids: setting conservation guidelines
Allendorf FW, Leary RF, Spruell P & Wenburg JK (2002) Trends in Ecology &
Evolution 16(11) 613-622
„rates of hybridization and introgression are increasing dramatically worldwide
because of translocations of organisms and habitat modifications by humans.
Hybridization has contributed to the extinction of many species through direct and
indirect means. However, recent studies have found that natural hybridization has
played an important role in the evolution of many plant and animal taxa.
Determining whether hybridization is natural or anthropogenic is crucial for
conservation, but is often difficult to achieve. Any policy that deals with hybrids
must be flexible and must recognize that nearly every situation involving
hybridization is different enough that general rules are not likely to be effective“
Hybridization
Natural
Type 1
Type 2
Natural hybrid
Natural
taxon
introgression
Type 3
Anthropogenic
Natural
hybrid zones
F1 only
Type 4
hybridization
without introgression
(from Allendorf et al 2000)
Hybrid swarm
Type 5
Type 6
Widespread
introgression
Complete
admixture
Recognize that if F1 hybrids are produced but there is no backcrossing,
and no F2 generation, then hybridization is irrelevant for both evolution,
and systematics. Thus, „hybridization“ is not neccessarily a problem
for the BSC, and F1 hybridization should
not be considered as evidence of the lack of reproductive isolation,
but rather simply the lack of a prezygotic barrier.
Prezygotic isolation: (prevents hybridization from occurring, but, how
stable are such mechanisms in a changing environment?)
Mechanisms preventing mating
geographic isolation (very common in freshwater fishes)
temporal isolation (very common in freshwater fishes)
different mate recognition systems (common within some groups, like
cichlids)
mechanical isolation (copulation unsuccessful, no transfer of gametes)
Mechanisms preventing fertilization
gamete mortality
genetic incompatibility (often when numbers of chromosomes differ, or
other chromosome level differences, there is also cyto-nuclear incompatibility)
Postzygotic isolation:
Mechanisms preventing development of interspecific hybrids
(thus the key issue of postzygotic mechanisms is to prevent hybridization)
zygote mortality (e.g. Dobzhansky-Muller incompatibilities)
hybrid inviability / sterility
METHODS OF DETECTING HYBRIDIZATION
Detecting natural hybridization between fishes is often complicated because several
situations are possible. Two closely related species, or conspecific populations, might occur in the same
habitat without ever interbreeding, yet hybridization may be suggested because of phenotypic overlap.
On the other hand, hybridization might occur, but the hybrids themselves may never breed for one of
several reasons, including infertility. If hybrids backcross with one or both paternal species,
introgression may have occurred or a hybrid swarm may be present. If hybridization is known to have
occurred, one may want to distinguish individuals of mixed ancestry from those of the two parental
species. However, as will be pointed out later, distinguishing individuals of mixed ancestry from
parental species is often impossible if hybridization has proceeded past the F1 generation.
Morphology
Hubbs & Kuronuma (1942) and Hubbs et al (1943) defined a statistic called the hybrid index in order
to measure the average morphological similarity of an individual fish to each of the two species or
other taxa. The index (I) is calculated separately for each character as I = 100 x [(u - X)/(Y-X)], where
u is the value of the trait for the individual being evaluated and X and Y are the mean values of the
trait for species X and Y. An individual fish with a value for the trait equal to X or Y will have an
index value equal to 0 or 100, respectively. An index value of 50 indicates exact intermediacy for the
character in question. The average value of the index for all diagnostic characters may be close to 50 in
both known and suspected hybrids, but the individual values may vary widely from character to
character (e. g. Ross & Cavender 1981).
Multivarite methods
Two major objections to the hybrid index are that (1) it requires the a priori indentification of
individuals from the two parental species and (2) it fails to account for the variances and covariances of
the discriminating traits. Highly correlated traits are often different measurements of the same
biological phenomena, reflecting the pleiotropic action of genes or the common response to
environmental effects (Falconer 1981).
Hubb's hybrid index has lost ground to multivariate statistical methods.
Multivariate statistical methods can circumvent these objections by deriving weighting factors for each
trait based on the variance-covariance structure of the data set.
Discriminant function analysis (DFA) derives weighted linear functions of the measured traits so that
two or more groups of individuals are maximally separated or discriminated in multivariate space.
DFA also requires a priori classification, but exclusion and inclusion probabilities for each indiviudal
are given. But the function can be distorted if two many individuals are at first, falsely classified. Thus,
the a priori classification should be done based on information independant of the variables used in the
function.
Principal components analysis (PCA) is a second multivariate statistical method that is often used in
hybridization studies. Unlike DFA, PCA does not require the a priori identification of groups or
individuals. PCA only derives a new set of uncorrelated variables, the principal components, which are
simple linear transformations of the original variables.
An example of a PCA plot (first vs. second component), based on
the morphology of two hybridizing fish taxa
PCA is relatively easy to apply,
and is an excellent exploratory
technique. In this example
involving a species that we work
with (Brachymystax lenok), 37
morphological characters were
used. While this researcher
(Sergey Alekseyev) did preclassify individuals (therefore the
circles are colored), it is clear that
he would not have had to. Also,
his technique is obviously good,
as he was also able to predict the
hybrids, which, appear to be F1‘s,
though no genetic data is used
here!
PCA does not, however, provide
probabilities of any classification,
but the components can be used
in subsequent analyses as input
variables into a large array of
other statistical techniques.
Here is another example of PCA used in the
analysis of hybridization.
Instead of a bivariate plot,
the first principal component is
simply plotted against the frequency of
individuals. Such a technique can be
informative when many of the original
traits (or at least those that are important)
are correlated, and
thus much of the variation in the
data set is contained within the first
component.
Figure 7.1 Frequency distributions of first
principal component scores for two cyprinid
fishes, Notropis cornutus and N. chrysocephalus,
and their suspected hybrids from each of five
geographic locations. From Dowling and Moore
(1984).
Common shiner
Notropis cornutus
Here is an example from our own
work, with grayling in Asia. After
phenotypic examination, as well as
the application of mtDNA typing, the
question arose as to whether two
sympatric populations were really
reproductively isolated. Examination
of morphological characters in
multvariate space show complete
distinction of the „Orange-spot“ and
„Lower Amur“ forms.
Note that neither PCA factor alone (nor any of the original variables) show nonoverlapping ranges, but considered together, there is clear separation. The additional
application of 10 nulcear markers (microsatellites) confirmed that there was no
evidence of any hybridization between these two forms, and thus, they could for
instance fulfill the BSC criteria of two distinct species.
Note as well, the morphological position of the third group (Upper Amur) which is
allopatric to the other two. What mechanism did you hear of in the last lecture that
might explain the morphological relationship of these three groups?
So you can see why the study of hybridization (i.e. hybrid zones) is closely linked to the
study of speciation. Hybird zones are model systems, where many of the mechanisms of
reproductive isolation can be investigated. The discovery of hybrids alone, however,
especially if they are only F1‘s , has little meaning beyond recognizing that a prezygotic
barrier is not existent. Post-zygotic mechanisms can be fully responsible (and often are)
for preserving the identity of two popuation groups (or species) where hybridization is
frequent. Prezygotic mechanisms may be more prone to breaking down under a changing
environment (or through anthropogenic influence) and thus the frequency of such
mechansims in freshwater fishes might parellel the frequency with which hybridization and
introgression is observed.
One additional scenario, common in some groups of freshwater fishes, is worthy of
demonstrating. Similar to the three-spine stickleback, parallel „speciation“ as been
proposed to occur among sympatric forms of many salmonid fishes, exemplified by the
genus Coregonus (found also in Salzkammergut!)
Sympatric dwarf and normal whitefish ecotypes
(Coregonus sp.)
Sympatric dwarf and normal whitefish ecotypes
(Coregonus sp.)
Such sympatric forms are often found in post-glacial lake systems
(that is, relatively young systems) whereby not only are F1 hybrids
found, but substantial degrees of introgression (i.e seen at genetic
markers). But the forms are real, and little phenotypic overlap is found.
What mechanism can maintain the existence of two such extreme forms
in the presence of substantial gene flow?
Detecting hybridization by genetic methods
Detecting natural hybridization and introgression by genetic methods is relatively straightforward
when the two parental species are completely fixed for different alleles at one or more loci. Individuals
of the two parental species will each be homozygous for different alleles, whereas F1 hybrids will be
heterozygous for these alleles at all diagnostic loci.
A DIAGNOSTIC LOCUS
Species 1
Locus 1
F1 Hybrid
Species 2
A/A (all individuals)
a/a (all individuals)
A/a
If hybridization has proceeded past the F1 stage, however, hybrid descendants will express a range of
different genotypes, including the two parental types. When populations are not „fixed“ for alternate
alleles, the problem is more complicated, and individuals can not be assigned to hybrid or pure classes.
Thus, hybridization indices must be quantitative.
To quantitatively demonstrate the likelihood of hybridization, Campton & Utter (1985) devised a
hybrid index measuring the relative probability that the composite genotype for each fish arose by
random mating within each of the two species. The index (IH) was defined as:
IH = 1.0 - log10 (px) / log10 (px) + log10 (py), where px and py are the conditional probabilities that the
composite genotype at all loci for an individual could have arisen by random mating within species X
and species Y, respectively, given the average allele frequencies for the two species and assuming
gametic phase (linkage) equilibrium among all loci.
The power of genetic methods to detect natural hybridization will increase as the number of
distinquishing loci increases. This is easily demonstrated by the following table.
Number of distinguishing loci
Probability of heterozygosity if one allele is found at 0.8 and the other at 0.2 across 1 to 6 different loci
Population
1
2
3
4
5
6
Parental (P)
0.3200
0.1024
0.0328
0.0105
0.0034
0.0011
F1 Hybrid (Ph)
0.6800
0.4624
0.3144
0.2138
0.1454
0.0989
43.3
92.1
Increased probability of a heterozygous hybrid
Ph/P
2.13
4.52
9.60
20.4
If the frequencies of alternate alleles (e.g. A and a) at a locus are 0.8 and 0.2 in one population and 0.2
and 0.8 in the second population, the probability of an individual being heterozygous in each
population is simply 2pq = 0.32. If the two populations interbreed, then the expected frequency of
heterozygotes among the F1 hybrids is 0.68, which is only 2.13 times the expected frequency of
heterozygotes in each of the parental populations. As the number of distinguishing loci increase,
however, this ratio also increases. For example, with four distinguishing loci, the frequency of fourlocus heterozygous among F1 hybrids is more than 20 times their expected frequency in each of the
parental populations (Table 7.1). This ratio increases to 92.1 with six distinguishing loci.
Thus, even without diagnostic markers, strong inferences about hybridization can be made, when a
number of loci are used.
For more hypervariable markers (like microsatellites), increasingly sophisticated statistical
techniques are available for estimating degrees of introgression (termed admixture) even when
allele frequencies of the parent populations are not known. However, there are also prior
assumptions and tradeoffs depending on the situation. The more information, the better. Some
aspect of the pedigree history is usually needed to make reliable inferences, (number of generations
since contact, prior alelle frequencies, number of hybrid categories, etc.).
Anderson EC & Thompson (2002) A method for indentifying species hybrids using
multilocus genetic data. Genetics 160, 1217-1229.
and other references therein.....
Mitochondrial DNA –
mtDNA can be an extremely helpful tool in understanding the dynamics of
hybridization and introgression, but only when applied in combination with
nuclear (i.e. bi-parentally inherited markers)!
That is, mtDNA helps to identify the source of the mother in a hybrid
situation, but alone, mtDNA can not distinguish between current hybridization
and ancient processes.
mitochondrial capture, or nuclear replacement, a consequence of
introgressive hybridization
X
nDNA
Y
mtDNA
nDNA
X
mtDNA
imagine a male of one
species (X), crossing with
a female of another (Y);
if we consider the black fill to represent the genome of the female,
both for nuclear and mtDNA, what will the F1 generation look like
in terms of the proportion of each parental donor?
nDNA
mtDNA
The F1 generation will have 100% mtDNA
of the female parent (Y), and a 50/50 mixture
of both parents (X + Y) for the nuclear genome.
in the next generation, a male X crosses with a hybrid female (a backcross)
nDNA
mtDNA
nDNA
mtDNA
X
what happens to the proportions of the parent genomes now?
the resulting backcross offspring still retains
100% mtDNA from species Y, but now
possesses only 25% of its nuclear genome.
With each similar backcross, approximately
½ of the remaining nuclear genome of species
Y will be lost and the result can quickly be
the complete nuclear „replacement“ or
„apparent“ mitochondrial capture of Y‘s
mtDNA into a population of X‘s nuclear gene
pool
Taxa
X
nDNA
Taxa
Y
mtDNA
nDNA
mtDNA
nDNA
mtDNA
X
nDNA
nDNA
mtDNA
nDNA
mtDNA
?
mtDNA
Thus, a small amount of interspecific gene flow
permits transmission of one species‘ mtDNA
genome into another species. Several mechanisms
can accelerate the rate of such an event. First,
an adaptive mutation in the mtDNA genome of the
female parent can lead to a „selective sweep“ and
ultimate fixation of one haplotype in two
hybridizing populations. Second, population
bottlenecks during glaciation and resulting genetic
drift could accelerate the rate of mtDNA
replacement.
Such events, like hybridization in general, can be
promoted through environmental change (which
changes the selection landscape), anthropogenic
actions such as the transport and stocking of one
species or strain, or any reason that reduces the
density of one of the species, promoting the chance
mating with another.
Studies using only mtDNA, can easily be misleading,
and such events may have occured a long time ago,
obscuring or confounding phylogenetic relationships.
My first tiger, isn‘t she pretty!
This is a photo of a so-called „tiger trout“, which is a cross between the brown trout (Salmo trutta) and
brook trout (Salvelinus fontinalis). Such crosses are produced in captivity, but the fertility of such
organisms in the wild is not well understood. The caption, came withthe photo, and demonstrates the
novelty of such hybrids among fisherman. This cross exists in the wild in Austria, and I have evidence
of introgession into wild populations of brown trout in the Tirol. Brook trout are introduced in
Austria, but also readily hybridize with native populations of Arctic charr (Salvelinus alpinnis).