Online Issues
<< All Back-issues
<< This Issue's Table of Contents
ILAR Journal V39(2/3) 1998
Comparative Gene Mapping
Jennifer A. Marshall Graves
| Jennifer A. Marshall Graves, Ph.D.. is Professor and Head of the School of Genetics and Human Variation, LaTrobe University, Melbourne, Australia. |
Comparative genomics--the cross-referencing of information on genome organization between species--provides an additional dimension to the Human Genome Project and can derive much information from it for the benefit of animal health and animal breeding. Arrangements of genes and other DNA sequences may be determined by a variety of genetic and physical techniques, at resolutions from the gross cytological level to the level of the single base pair. Gross arrangements and rearrangements can also be charted by comparative chromosome painting. Genome organization may then be compared across mammal--and other vertebrate--species. Genetic mapping is well advanced in several livestock species as well as rodent model species, and outline maps are available for at least 30 mammal species in 8 orders. At the time of this writing, maps are being rapidly constructed for chicken and fish species. Comparisons, even over vast evolutionary time scales, show that the mammal genome--indeed, the vertebrate genome--has been highly conserved. Thus information about location and function of genes is directly transferable across species and should greatly speed up the search for genes that specify inherited diseases in domestic mammals and humans as well as genes that specify economically important traits.
INTRODUCTION
The map and base sequence of the human genome will stand as a monument to 20th century technical innovation and doggedness. A detailed gene map is already available, physical maps encompass several chromosome regions and 2 whole chromosomes, and sequencing is accelerating toward a scheduled completion date of 2003, 2 yr earlier than originally targeted. At the time of this writing, the human map is already being used for locating genes that act to bring about variation in outward characteristics and disease states (phenotypes).
The prodigious output of the Human Genome Project can be directly applied to increase our knowledge of, and control over, every mammalian species of use and interest to humans and can even be applied to other vertebrates such as birds and fishes. Comparative genomics can also use knowledge gained from other animal species to advance the goals of the Human Genome Project and to contribute to an even broader goal--understanding the evolution of the human species.
GENES AND GENOMES
The genome is defined as the full set of genes that defines an organism. Mammals and other vertebrates, being diploid, have 2 full genomes in each cell, 1 from the sperm and 1 from the egg.
Genome Function and Organization
Oddly, considering their unsuitability as experimental organisms, humans are our mammal type-species. The human genome is by far the best known genome of any mammal and so is the obvious point of all comparisons. It contains approximately 70,000 genes, of which about 10% have been identified and mapped and another 20% have been ordered as short (100-base pair) runs of identifiable unique sequence. These genes have functions in basic metabolism or in determining specialized characteristics of the thousands of different cell types that make up the human body. The protein product of a gene may act as an enzyme (or an enzyme subunit), a transport molecule, a structural protein, a hormone, or a regulator in any of the myriad biochemical reactions of the body.
Each gene consists of a sequence of DNA. which is transcribed into RNA in the appropriate tissue or organ, then translated into a single polypeptide in the cytoplasm of cells. In this process, the sequence of nucleotides (bases) along the double helical DNA molecule is translated through the fabled genetic code in which each amino acid in the polypeptide product is specified by a triplet of bases. The average length of coding portion of genes is about 1000 base pairs (1000 base pairs = 1 kilobase, or kb). Between the genes are long stretches of DNA often full of repetitive sequences, which have no obvious function and may simply represent genetic junk. Even within genes, intervening sequences of noncoding DNA ("introns") are interspersed between coding regions ("exons"), and are spliced out of messenger RNA after transcription. Overall, only about 3% of the mammalian genome is coding DNA.
In all organisms, genes are parts of immensely long DNA molecules. Thus genes are arrayed linearly along chromosomes, and this linear array can be represented its a linear map.
The entire mammalian genome, which represents more than 1 m of DNA, is divided up into smaller lengths for ease of handling during cell division, when the genome is replicated and the daughter DNA molecules are segregated. These DNA molecules constitute the chromosomes that we can see as staining bodies when they are coiled and packaged up with protein, ready to be apportioned at mitosis.
Chromosomes may be photographed or imaged at mitosis, then arranged as a karyotype in order of descending size. Because vertebrates are diploid, there are 2 copies of each chromosome. Chromosomes are distinguished, as well as by their size, by the position of the centromere, which remains undivided, attaches spindle fibers at mitosis, and appears as a constriction. Chromosomes may also be linearly differentiated by several methods (such as Giemsa or G-banding1) that induce specific patterns of dark and light bands along their length. This pattern reflects the underlying sequence organization, and G-band patterns are therefore diagnostic for each chromosome and consistent within the species.
Genome Comparisons
In all mammals, the genome is basically the same size, comprising approximately 3300 million base pairs (or Megabases [Mb]) and about the same number of genes (approximately 70,000). These genes represent much the same set as in humans, performing the same housekeeping and specialized tasks. As for the human genome, other mammalian genomes are composed largely of noncoding, mostly repetitive sequences. Compared with those of mammals, the genomes of birds are only about 1/3 the size, and the genome of the pufferfish (Fugu) is especially compact at about 1/8 the mammal genome size, although it appears to share a similar set of genes.
The equivalent genome is divided differently in different mammal species. Haploid chromosome numbers (that is, the number in a single genome) range from n=3 large chromosomes in the Indian muntjac to n=67 small chromosomes in the black rhinoceros. This karyotypic variation--the different numbers, sizes, and shapes of chromosomes in different species and the difficulties of comparing G-band patterns--for a long time led to a belief that many rearrangements had scrambled gene orders beyond recognition in different mammal lineages. However, comparative mapping over the last 2 decades has shown that the mammalian genome is much more conserved than was apparent from cytogenetic comparisons, and this conclusion is now being reinforced by direct observations of comparative chromosome painting. The remarkable conservation of the mammal genome--indeed, the vertebrate genome--has resulted in the ready transfer of information between different mammal species and the belief that deducing the shape of an ancestral mammalian genome--maybe even an ancestral vertebrate genome--is not a vain hope.
Comparative genomics includes comparisons of every type of gene map at every level from the cytological to the molecular. These methods are described and explained below in Markers for Genome Analysis. The text summarizes the value of such comparisons in practical outputs for animal health and animal breeding as well as for using animals as models for human disease. The summary also describes our first attempts to solve the great evolutionary jigsaw puzzle--to deduce the ancestral arrangements from which the genomes of all living mammals derived.
MARKERS FOR GENOME ANALYSIS
Different strategies may be used to detect genes and DNA sequences and to compare their positions and arrangement in different species. Classically, genes were recognized by their effects on the phenotype, and a genetic map was constructed by observing the patterns of segregation among the offspring of parents differing in 2 or more characteristics. Genetic maps still rely on inherited differences, but these may be recognized at the molecular as well as the phenotypic level. Physical maps describe the physical location of a gene or a DNA sequence on a chromosome or chromosome region, with greater or lesser resolution. Maps on a molecular scale can be constructed with the base sequence of the region at their endpoint.
Different types of maps require different types of gene marker. Genetic (linkage) mapping relies on intraspecies variation (alleles) at a particular gene locus, whereas somatic cell genetic mapping requires interspecific variation, and physical mapping techniques require a cloned DNA probe. The availability of molecular and phenotypic differences between alleles has vastly increased the ease of genetic mapping, especially in humans, where information is compiled from limited and uncontrolled crosses. Importantly, this availability has also made it easy to line up genetic and physical maps of the same species and to move between them in physically isolating a target gene.
Molecular Markers
Molecular markers have been developed as the result of gene cloning--from the ability to isolate and replicate small pieces of large genomes by splicing them into small self-replicating DNA molecules of bacteria or virus vectors. Vectors have different characteristics that make them suitable for different types of study. At the lower end are lambda virus vectors, which accept 15 to 20 kb of insert DNA, and bacterial plasmids and cosmids, which can accommodate slightly more. At the upper end are bacterial and yeast artificial chromosomes (BACs1 and YACs1), constructed by stringing together the components of normal chromosomes, which may accept pieces of foreign DNA up to a few megabases. The entire genome of a mammal can be fragmented into appropriate sized pieces, which are inserted randomly into vectors and propagated as mixtures (libraries) of colonies or plaques on a Petri dish. The vector carrying a particular sequence can then be isolated by cloning the progeny of a single virus, bacterium, or yeast.
One variation is to start with messenger RNA prepared from a particular tissue, make DNA copies (cDNA) and splice them into a vector. These cDNA libraries provide intronless copies of all the genes active in a particular tissue, so they enable us to concentrate on the minority of the genome that codes for protein. Very rapid progress has been made in identifying all the expressed genes in the human genome as minimally sequenced expressed sequence tags ("ESTs") whose unique position can be determined on the map.
An alternative for isolating a single sequence from a complex genome is to replicate it selectively, exploiting the predilections of the DNA replication system in the polymerase chain reaction (PCRI). Under in vitro conditions, DNA polymerase will replicate DNA only when provided with 2 short (approximately 24 bases) "primer" sequences, which are constructed to be complementary to short sequences on either side of the target. Over a number of replication cycles, PCR will therefore amplify a target sequence between 2 primers in a chain reaction.
Variation in DNA markers can be detected in a number of ways. Base differences can be identified as restriction fragment length polymorphisms (RFLPs1) if they happen to fall within a recognition site for a restriction enzyme. Otherwise, they may be detected by hybridizing with allele-specific oligonucleotides, or allele-specific primers for PCR, which bind only if there is a perfect match. Single base differences in a DNA molecule can also be detected by single strand conformation polymorphism.
Especially variable and therefore exceedingly useful markers have been repetitive sequences, which frequently differ in the number of tandem repeats. Initially, variable numbers of tandem repeats (or minisatellites) were detected by hybridizing with the (often lengthy) core sequence and detecting different sized bands on a Southern blot. A very sensitive detection system is now available to detect variation in the numbers of very simple sequence repeats (di-, tri-, or tetranucleotides) by PCR from primers lying outside the repeat stretch. The product length is assessed on an acrylamide gel that can separate alleles differing in only I or a few repeats. These simple sequence length polymorphisms, commonly known as microsatellites, have revolutionized genetic analysis of complex genomes.
Type I and Type II Markers
Many of the markers most useful for constructing a detailed linkage map are anonymous DNA sequences, most of which are derived from the majority noncoding and therefore more variable class of DNA (type I markers). However, these highly polymorphic markers are of very limited use for comparisons between genomes because their variability makes it impossible to detect homology across species. Although they may be employed in linkage mapping in closely related species (for example, cattle microsatellites have been valuable for constructing a sheep linkage map; Broad and others 1998), they are not likely to recognize homologous sequences between more distantly related species. The most useful markers for comparative genetics are the highly conserved coding genes (type II markers), which may show greater than 90% sequence homology within the exons between all mammals, even all vertebrates, and be recognizable even in Drosophila and yeast.
Markers for mapping coding genes must be found in variation of the properties of a gene, its protein product, or the resulting phenotype. Changes in the gene product are usually detected electrophoretically as altered charge on the protein. Changes in the base sequence of the gene itself can be detected as RFLPs or simple sequence length polymorphisms (microsatellites) within the gene. The problem with type II markers is that the very conservation of coding genes means that there are few polymorphisms available for placement on the genetic map. A solution is to screen introns for sequence differences or microsatellites that can be recognized by amplifying from conserved exon sequences. Sets of conserved "universal" primers (Venta and others 1996), or comparative anchor-tagged ("CAT") sequences (Lyons and others 1997), made by designing primers to exon sequences conserved between at least 2 species, have been of considerable use for comparative mapping, at least in some mammalian species, although many "universal" primers have proved to have limited applicability.
A particularly useful strategy to increase the supply of polymorphic markers has been to use hybrids between closely related species. This has been spectacularly successful in constructing a detailed linkage map of the mouse genome using Mus musculus x Mus spretus crosses (Davisson and others 1998), but the same idea lies behind the development of crosses between bovine species, deer species, cat species, and subspecies of the tammar wallaby. The same strategy was used 20 yr ago to map some fish species (Morizot and others 1998).
Criteria of Homology and Nomenclature Problems
Genes often belong to large gene families with similar sequences. A big problem for comparative mapping is distinguishing between orthologues (homologous genes in different species, such as alpha globin in human and mouse) and paralogues (genes within the same species descended from the same ancestral gene by duplication and divergence, such as human alpha and beta globin). It has therefore been necessary to develop specific criteria by which orthologues may be recognized in different species at the DNA, protein, or phenotype level (CGOW 1996). The most stringent of these criteria are
- at the nucleotide level, similar nucleotide sequence or at least cross-hybridization to the same DNA probe. Similar transcription profile is helpful supplementary data but cannot be regarded as diagnostic because there are too many exceptions.
- at the protein level, similar amino acid sequence. Similar biochemical properties, expression profile, and immunological cross-reaction are regarded as helpful but less robust.
- at the phenotype level, complementation of function is a much superior criterion than similarity of mutant phenotype.
It is a sign of the recognition of the extent of genome conservation that an important criterion of homology has now become conserved map position, which can often be used to distinguish between paralogues. Authors must always state the criteria by which homologues have been recognized.
Gene nomenclature has been a continuing problem in comparative gene mapping, although this has been mitigated by agreements by most species groups (even rat) to adopt human nomenclature where homology is clear. Mouse nomenclature still retains symbols originally descriptive of mutant phenotypes. Although this practice is too well established to make changing practicable, the same symbol is now used where possible for new genes described in human and mouse.
Species designation also remains very inconsistent in comparative genomics publications. The convention recommended by the Comparative Genome Organization Workshop is a 4-letter code derived from the species name and italicized to reflect standard species nomenclature. Examples include Hsap for Homo sapiens, Mmus for M. musculus, Meug for the tammar wallaby (Macropus eugenii), and Ggal for the chicken (Gallus gallus).
Terms particularly useful for comparative gene mapping, as defined by the Comparative Genome Organization Workshop (1996) are
- homology--regions containing arrays of homologous genes in different species;
- conserved synteny--syntenic associations of the same genes in different species, regardless of gene order or interposition of other genes;
- conserved segment syntenic associations of contiguous genes in different species. The shortest conserved segment in different species has been called a SCEUS, for "smallest conserved evolutionary unit segment"; and
- conserved order--3 or more homologous genes in the same order in different species.
MAPPING METHODS
Two types of map describe a genome: genetic maps, which are derived from recombination frequencies between genetic markers; and physical maps, which are constructed from information about the physical location of genes on chromosomes. The 2 maps may be aligned if they share markers.
Genetic Maps
Genetic maps were devised in the early 1900s by Thomas Hunt Morgan and his students, using crosses in fruit flies. This discovery followed observations in the early 1900s that some genes acted as if they were "linked" together in the offspring, whereas others segregated independently, according to Mendel's laws. Linkage was explained by the tendency of 2 markers on the same parental chromosome to be passed on together, in contrast to markers on different chromosomes, or far apart on the same chromosome, behaving independently at meiosis. Linkage reflects the physical reality of crossing over between 2 homologous chromosomes at meiosis. Morgan observed that the amount of recombination between 2 gene loci varied widely, and his student guessed that the recombination percentage between 2 loci was related to how far apart they were physically.
The technique for genetic mapping is therefore to mate parents differing in 2 or more traits (that is, having different alleles at 2 or more loci) and then to score the patterns of segregation of parental alleles among the offspring. 2 genes on different chromosomes will randomly segregate at meiosis, so that 1/4 of the offspring will show each of 4 combinations of alleles at the 2 loci. However, if genes are close together on the same chromosome, parental combinations of alleles will pass to the offspring together, unless they are separated by recombination. Since the probability of recombination increases as the physical distance between genes, the recombinant percentage progeny provides a measure of the genes' relative distance apart. The unit of measurement is called a centiMorgan (cM) after Morgan. For markers close to each other (where the complication of double recombination can be ignored) 1 cM = 1% recombination.
If the parents differ at 3 or more loci, recombination between them is additive, allowing for multiple recombination. This allows the 3 loci to be placed in a linear array in which recombination percentages represent the relative distances between the loci in this "linkage group." A map can be built up in stages from different crosses in different laboratories, and ultimately, the linkage group describes the gene order along a whole chromosome.
Linkage maps depend utterly on the availability of polymorphisms--the existence of 2 or more alleles at a locus. This was traditionally the great limitation of mapping, but the availability of highly polymorphic DNA markers (RFLPs and especially microsatellites) has revolutionized genetic mapping in many animal species.
Physical Maps
Genes may be assigned to physical positions within chromosomes or chromosome regions by somatic cell genetics, radiation hybrid mapping, and in situ hybridization. A molecular description of the genome, given by restriction mapping and nucleotide sequencing, also qualifies as physical mapping, but is considered below separately.
Somatic cell genetics was developed in the 1970s and in situ hybridization in the 1980s. Neither method requires a breeding colony, or even a live animal, as long as cell samples are available. Nor do they require polymorphic markers, since somatic cell genetics uses interspecific variation and in situ hybridization requires only a cloned sequence big enough to produce a signal. However, both methods represent low resolution techniques. Somatic cell genetics simply establishes synteny groups and assigns them to a chromosome, but does not specify position or order; and in situ hybridization provides a rough cytological localization on a chromosome.
Somatic cell genetic analysis uses viable hybrid cells derived from fusion of somatic cells from different species. It depends on the observation that chromosomes are lost from only 1 of the 2 parental sets (at the time of this writing, we still do not know the reason). For example, rodent-human hybrids segregate human chromosomes, so that it is possible to derive a hybrid panel uniquely representing each human chromosome. Hybrids all retain and express the full set of rodent genes; however, a hybrid will retain and express only the human genes on the particular human chromosomes retained in that hybrid. Thus, by detecting patterns of presence and absence of human markers in a set of hybrids and correlating these with the patterns of the presence or absence of particular chromosomes, it is possible to assign a human gene to a particular human chromosome. Some regional mapping is possible using hybrids that retain only portions of a chromosome.
The method described above was developed particularly for humans but has been used extensively since the early 1980s to assign genes to chromosomes in cattle, cat, and marsupial species. Hybrid panels have also been developed for more exotic species such as mink, arctic fox, and vole (Rubstsov 1998; Serov 1998) and most recently for shrew (Nesterova and others 1998).
Radiation hybrid mapping is an extension of regional mapping. Hybrids are constructed from donor cells (cells from the species to be mapped or cell hybrids bearing a single chromosome of the species to be mapped) that have been lethally irradiated to cause chromosome fragmentation. Hybrids therefore contain only small regions of the irradiated donor genome incorporated into chromosomes of the unirradiated parent. These radiation hybrids are more likely to bear 2 genes if they are physically close together on a chromosome, so the frequency of concordance of markers may be used as a measure of their physical proximity.
In situ hybridization is a method by which a cloned probe labeled with radioactive isotope or fluorescent tag is bound specifically to the DNA sequence to which it is complementary, within the framework of the chromosome fixed to a microscope slide. Radioactive signal is detected by autoradiography. Fluorescent tag is bound indirectly to the probe by layers of specific antibodies that detect molecules bound to the DNA (for example, biotinylated probe is bound to avidin and detected by fluorescent antibodies). Fluorescence is detected by a sensitive ultraviolet light microscope. Fluorescence in situ hybridization (FISH1) is remarkably sensitive as long as the probe is homologous and long and the background of repetitive sequence is suppressed by competing with unlabeled whole DNA or the repetitive fraction. It provides a localization to a region about 1% of the genome length (that is, approximately 30 Mb). Different fluorescent dyes produce signal at different wavelengths so that if the efficiency is high enough, 2 or more colors may be used to identify different sequences within the same cell.
Chromosome Painting
Chromosome painting is a fluorescence in situ hybridization technique that differs from FISH in that it uses a unique DNA probe derived from a whole chromosome or chromosome region. Chromosomes from a species may be physically separated by flow sorting or microdissection, as described by Ferguson-Smith and others (1998). DNA from a single chromosome may then be PCR amplified using degenerate oligonucleotide primers so that all sequences are represented. When a single chromosome paint is applied to chromosome preparations from the same species under suppression hybridization conditions (so that repetitive sequences shared between many chromosomes are not detected), only the 2 copies of that chromosome are hybridized. The regions that hybridize to the paint are then detected by a fluorescent tag (in the same way as for FISH) and appear as a colored region.
Paints have been prepared from each of the flow-sorted human chromosomes. In the same way, single chromosome paints have been prepared from all the chromosomes of the cat, mouse and several farm mammals as well as marsupials and chicken (Ferguson-Smith and others 1998). Paints can also be prepared from regions of chromosomes--even single G-bands--by microdissection.
A single chromosome paint from I species may then be applied to chromosome preparations of another species under suppression hybridization conditions so that it binds only to homologous regions. A pattern of regions homologous between species may be obtained. Again, different dyes may be used to produce signal at different wavelengths. Different combinations of 3 dyes can produce 24 distinguishable signals so that painting with all the human chromosomes simultaneously may be performed (Wienberg and Stanyon 1998).
Comparative chromosome painting (or ZOO-FISH) has been most effective when performed between species reasonably closely related such as human and apes (Wienberg and Stanyon 1998), mouse and rat (Ferguson-Smith and others 1998), or 2 species of kangaroo (Toder and others 1998). Good signal has also been produced by hybridizing human paints onto carnivore, ungulate, and even insectivore chromosomes (summarized in Glas and others 1998). Painting rodents with human paints has been more of a challenge because there have been many more rearrangements as well as more sequence divergence.
Painting is an extremely direct way to assess the amounts of rearrangement between 2 species, with the advantage over comparative mapping of providing direct information on homologies over an entire genome in about 1 wk. Such information would otherwise require a relatively detailed comparative map, taking many years to construct. Chromosome painting can rapidly extend results from comparative mapping. For instance, the cat map covers 50 to 60% of the genome and the painting, more like 90%.
However, the resolution of the method is a limitation. Painting between relatively closely related species can detect rearrangements as small as 5 to 10 Mb (Wienberg and Stanyon 1998). Resolution can be improved by using reciprocal painting between 2 species to ensure that small unpainted regions, which are difficult to detect in a brightly fluorescing background, are not overlooked. The inability to detect rearrangements within conserved chromosome blocks is also a limitation.
Mapping at the Molecular Level
Physical mapping using somatic cell genetic or in situ hybridization techniques has limited resolution. At the other end of the scale are methods that provide physical information at a molecular level: restriction mapping of cloned sequences, contig construction and, ultimately, complete base sequencing.
Restriction mapping of a region of DNA is based on the recognition of specific base sequences (usually 4- or 6-base palindromes) by restriction enzymes, liberating restriction fragments of specific sizes that can be visualized after separation by gel electrophoresis. Different enzymes cleave the same DNA into overlapping fragments. The arrangement of the restriction sites can be deduced by comparing the fragments released by digestion with restriction enzymes singly and in combination. Long range restriction maps can also be produced using enzymes that cut infrequently at 8-base, 10-base, or even longer recognition sites. In this way, cloned pieces of DNA (small lambda clones up to megabase-sized YACs) may be mapped and compared. Very detailed patterns of cut sites can be recognized in overlapping clones, so a map of contiguous regions of DNA (contigs) can be constructed.
The ultimate information about a region of the genome is its nucleotide sequence. Sequencing of fragments (usually subloned restriction fragments or target sequences between PCR primers) of a genome is relatively straightforward, although it is still a massive undertaking for more than just a gene and its surroundings. The ultimate goal of the Human Genome Project--to sequence the entire genome by 2003--appears realistic in view of accelerating progress, with large tracts of several human chromosomes sequenced.
However, for other mammalian genomes, sequencing is a rather inefficient way of obtaining information about coding regions, since a majority of the genome is noncoding, and there is only I gene on the average every 40 to 50 kb. Thus, mass sequencing is unlikely to be undertaken for any other mammal except mouse, in which a sequence-ready physical map is already being constructed at the time of this writing (Davisson and others 1998). Sequence comparisons in mammals are therefore usually made on a gene-by-gene basis.
Mass sequencing of random clones is practicable for the compact genome of the pufferfish Fugu, which has a gene density of I per 6 kb. Since breeding this species is difficult, sequence scanning is the method of choice, and direct scanning of (usually human or mouse) databases is the method of comparison with mammals (Elgar and Clark 1998).
Putting the Maps Together
The potential for cross-referencing the different types of maps is important because it allows us to move from a genetic map (which may contain markers known only as phenotypes, like many genetic diseases) to a physical map and ultimately to a molecular map. This cross-referencing provides the tools to identify and physically isolate the gene. What limits this comparison is the type of marker from which the map is built up; genetic maps are composed largely of highly polymorphic DNA markers (type I), whereas physical maps are composed largely of cloned genes (type II). It is therefore wise to include coding genes within genetic maps so that they can be transferred to physical maps and compared across species.
A physical description of the DNA molecule that constitutes a chromosome--or an entire genome--can be described at several levels of resolution. At the cytological level, some detail (to approximately 10 Mb) is given by medium-resolution G-banding and chromosome painting. A low level of resolution of gene location is provided by somatic cell genetics and in situ hybridization (approximately 30 Mb), but linkage mapping can resolve markers as close as 1 cM, which is equivalent to 1 to 2 Mb. The molecular level represents the other extreme level of resolution. Restriction maps provide a linear array of restriction enzyme cut sites, usually several per kilobase over distances of tens of kilobases. The ultimate in resolution is the base sequence.
A gap exists between the resolution offered by genetic and physical mapping and molecular characterization. This gap can now be filled by large insert clones (BACs and YACs), and contigs built up by their overlap that span up to whole chromosomes. Radiation hybrids also provide an intermediate level of resolution where, at least in human, radiation hybrid mapping has permitted the physical ordering of several thousand human loci. This method offers the potential to order large or small chromosome segments isolated in a somatic cell hybrid, potentially providing the framework for building a YAC contig.
DEPTH AND BREADTH OF INFORMATION
All these methods have been used to various extents, depending on the resources available, to create gene maps of different mammal species. Some species (such as mouse) are inexpensive to keep in colonies and easy to breed, essential factors for constructing a linkage map. For some, there is the added benefit of available interspecies crosses to maximize polymorphisms. Other species (such as lions and whales) would not be so amenable to linkage analysis. Somatic cell genetics has been particularly useful for mapping genes to human chromosomes because of the characteristics of rodent-human hybrids, which rapidly segregate human chromosomes. Primate genes have been mapped by the same strategy, as well as bovine and cat genes and genes on the marsupial X. However, rodent genes are a challenge because most hybrids perversely retain a full set of rodent chromosomes. Likewise, in situ hybridization has been most useful for species that have easily distinguished chromosomes (such as human, hamster, and various marsupial species) and for which many cloned genes are available.
Progress in Constructing Linkage Maps
At the time of this writing, very detailed linkage maps (with an average distance between markers of less than 0.01 cM) are constructed for mouse, with about 6000 genes and 13,000 DNA markers mapped, largely through the availability of interspecies crosses and backcrosses (Davisson and others 1998). A linkage map for the rat has been started more recently but now contains 900 known genes and about 4000 microsatellites (Levan and others 1998). Major, mutually supportive efforts to construct good linkage maps for livestock species have resulted in the rapid expansion of maps of cattle, sheep, and pig. A medium density linkage map (more than 1400 markers) has been constructed of the bovine genome (Womack 1998). A map of the sheep genome with 519 markers (mostly microsatellites) has been made; a 4.2-cM third generation map of the sheep genome is under construction (Broad and others 1998). The horse map is progressing rapidly, preliminarily with 150 markers (Bailey and Binns 1998). The dog map has been started with more than 100 microsatellite markers, but so far few coding genes (Binns and others 1998), and a 5-cM cat map has been derived from interspecies crosses (O'Brien and others 1988).
Linkage mapping in other species has begun. A 10-cM map of the baboon has recently been completed with 330 markers (Rogers and VandeBerg 1998). A framework linkage map is now available for the marsupial Monodelphis domestica with 69 loci, half of them anonymous DNA markers (Samollow and Graves 1998).
Three chicken linkage maps at average densities of 6 to 9 cM containing up to 643 markers and 80 phenotypes have been rapidly produced using reference crosses from Europe and the United States (Burr and Cheng 1998). The growth of linkage maps for different fish species has been even more remarkable due to the special characteristics in some species (for example, haploid progeny in zebrafish make backcrosses unnecessary). Approximately 650 markers have been identified on the zebrafish map, including 100 coding genes (Postlethwaite and others 1998), 334 markers ( 103 genes) on the Xiphophorous map and 170 markers on the medaka map (Morizot and others 1998). The beginnings of linkage maps have been sketched out for several commercial species including trout and salmon.
The total size of the human genome is approximately 3700 cM, and that of other mammals is comparable. For example, the mouse map is 1400 cM, the dog map is 2100 cM, and the sheep linkage map is 3500 cM. The chicken map is approximately 3800 cM, although the genome is physically only about 1/3 the length of the mammalian genome. The variation is more likely to reflect differences in rates of recombination rather than large discrepancies in genome size.
Progress in Physical Mapping
Somatic cell genetics revolutionized human gene mapping, establishing a framework for autosomal gene maps for the first time. It also enabled gene mapping in great ape and other primate species, for which no suitable breeding colonies existed. It provided the first autosomal maps for many other species (such as bovine and cat; O'Brien and others 1998; Womack 1998) and is still used to build up the outline of a physical map, even when higher resolution maps are available (such as sheep; Broad and others 1998). More recently, somatic cell genetic analysis has provided a rapid and relatively low cost means to assign largely type I markers to chromosomes as a start to establishing horse and dog linkage maps (Bailey and Binns 1998; Binns and others 1998).
Somatic cell genetic mapping is still the method of choice for rapidly (and relatively cheaply) establishing a framework map for species like the shrew or the vole (Nesterova and others 1998; Serov and others 1998), in which no polymorphisms are available and breeding has been difficult. The method has also been critical for building up maps for mink (all 77 genes), fox (all 35 genes), and marsupials (the first 20 genes in the tammar wallaby) (Rubstov 1998; Samollow and Graves, 1998; Serov 1998). However, the unstable karyotype and the chromosome fragmentation in cell hybrids produced between rodents and marsupials, monotremes and birds has rendered the method less than ideal for mapping in these groups, although other species combinations (such as vole x M. domestica) appear more tractable for these distantly related species (Nesterova and others 1997).
Radiation hybrid mapping also has the advantage of requiring no intraspecific variation or breeding populations, with the potential to order type I or II markers within small regions. Radiation hybrid panels are now available for human, baboon (Rogers and VandeBerg 1998), bovine, pig, mouse (Davisson and others 1998), chicken (Burt and Cheng 1998), and zebrafish and are under way for many other species.
In situ hybridization is used in a growing number of species and is most useful where more accurate linkage or molecular mapping methods are unavailable. It requires the availability of DNA clones, so is useful only in species in which DNA libraries are available and at least a few genes have been cloned. It is possible to use the low resolution, but very sensitive, radioactive in situ hybridization to localize heterologous clones, even from species as distant as human and marsupials, monotremes, or chicken, as long as very conserved genes are used (such as the human Duchenne muscular dystrophy probe). Most of the gene assignments in monotremes have relied on radioactive in situ hybridization using human cDNA probes to conserved genes (Graves 1998).
Advances in fluorescence in situ hybridization have resulted in the greater sensitivity of this technique. FISH offers better efficiency and resolution but requires long and almost 100% homologous DNA probes. However, in the great apes, FISH with large insert human probes (cosmids and YACs) has proved to be very effective. FISH has played a critical role in anchoring unknown synteny or linkage groups to a particular chromosome in many species. For example, 85 bovine linkage groups were located on 26 chromosomes by FISH mapping 1 or more representatives, and the last of the unknown bovine synteny groups finally found a home when I of the genes was located by FISH. With improved methods for identifying chromosomes that are morphologically very similar, more than 100 FISH localizations now exist in cattle (Womack 1998), and the method has become useful even for sheep.
Molecular mapping is far advanced in some species and quite impracticable in others. Genomic and cDNA libraries in phage, plasmid, and cosmid vectors have been constructed from model mammals and livestock species, and even from more exotic mammals like wallaby and platypus. Chromosome-specific libraries, constructed from DNA of sorted chromosomes, are being developed for cattle (Womack 1998). Gridded YAC and BAC libraries are available for cattle, pig, and sheep (Broad and others 1998) and BAC libraries, for dog (Binns and others 1998).
Conserved Synteny in Mammalian Gene Maps
Early comparisons at the cytogenetic level painted a nihilistic picture of complete genome scrambling between mammal groups. However, even the first comparative gene map-ping--particularly between cat, bovine, and human--showed a level of conservation that could not be appreciated using G-band comparisons (reviewed in O'Brien and others 1988). The ensuing 2 decades of intensive mapping of hundreds of loci across more than 40 species has consolidated a picture of quite extraordinary genome conservation.
A giant jigsaw of more than 900 genes mapped in 32 species, constructed both from the data available in this issue and in databases listed by Wakefield (1998), is presented as the poster entitled "Comparative Genome Maps of Vertebrates'' and enclosed in this issue (Wakefield and Graves 1998). Very large regions of conserved synteny with the human genome are apparent when genes on the same chromosome are joined by vertical lines. For instance, most of human chromosome 9 is represented by cat D4 and pig 15, and human chromosome 17 is represented by mink 5, pig 12, and bovine 14. Of the 23 human chromosomes, 16 are represented by a single cat chromosome, and the other 7 are split between 2 cat chromosomes, representing a total of 30 homology segments if internal rearrangements are ignored (O'Brien and others 1997). It is evident from this comparative map that large autosomal regions have been conserved between human and primates, carnivores, artiodactyls, rodents, and insectivores.
Of all chromosomes, the X stands out as almost completely conserved between different eutherian mammals. Of the hundreds of genes on the human X that have been mapped in 1 or more of 37 other eutherian species, all are on the X except 3 exceptional genes that map to the X in human and an autosome in mouse. The exceptional conservation of the X was recognized decades ago and was proposed to be the result of selection against disruption of the chromosome-wide X inactivation system (Ohno 1967).
Some exceptions to the picture of conservation of synteny appear inconsistent with the evolutionary distance between species. For instance, within the primates, which are generally very conserved compared with the human genome, the gibbon stands out as having multiple breaks in synteny. In addition, even with the few dog genes that can be compared, it is evident that the dog map is more fragmented compared with human than is the highly conserved cat genome.
Rodent maps are much more broken up with respect to human maps. Even the earliest maps of mutants and isozyme loci indicated that the conserved segments between the mouse map and the human map are small. Detailed comparisons and subsequent analysis led to the conclusion that there have been 150 rearrangements between the species, leaving the average length of conserved segments as only 8.1 cM (Nadeau and Taylor 1984). This limited conservation is even more obvious now that the maps of both species are so detailed and is apparent from the comparative mapping poster, even though not all of the mouse loci are represented (Wakefield and Graves 1998).
It was initially thought that this breakdown of synteny merely reflected the increased evolutionary distance between rodents and primates. However, the hamster genome shows considerably more synteny with human than does the mouse, even though caviomorphs probably diverged from the primate lineage even earlier than rodents, and the same appears to be the case for the even more distantly related shrew (Serov and others 1998) and even chicken (Burr and Cheng 1998). For instance, human chromosome 12 is broken up in mouse and rat but is intact in the chicken genome, suggesting that the mouse has a particularly rearranged karyotype that is not typical of other orders and perhaps not even of other rodents.
Conserved Synteny in Distantly Related Mammals and Other Vertebrates
Although marsupials and monotremes have genome sizes in the range of eutherians, their karyotypes are very distinctive, and it was initially expected that their genome arrangements would be scrambled beyond recognition, by comparison with the human genome. However, this is not the case. For example, 7 genes spanning human chromosome 17 all map to linkage group 3 in M. domestica, and 7 human chromosome 3p genes lie on chromosome 2q in the wallaby (Samollow and Graves 1998). Even in monotremes, the mammals most distantly related to humans, conserved synteny is observed, particularly in the X chromosome, which shows the same distribution of most X-linked and autosomal locations (Graves 1998).
Remarkably, conserved synteny is apparent, at least in short regions, even in much more distantly related vertebrates. Birds, which evolved from a branch of reptiles that diverged from mammals 350 million yr ago (MYA1), show many shared syntenies of coding genes. For example, 8 human chromosome 6 genes all map to chick chromosome 3 (Burt and Cheng 1998). Even more remarkable is the conservation of synteny apparent in comparisons between humans and fishes; for example, 10 markers on human chromosome 2q all lie within zebrafish linkage group 9 (Postlethwaite and others 1998). Conservation of gene arrangement is all the more remarkable in fish that have genomes only a fraction of the size of the human genome. They contain the same genes, having the same structure and similar sequence, and may even be in the same order, although interrupted by much less repetitive DNA. Vertebrates such as birds (with a genome 1/3 the size of mammal genomes) and especially pufferfish (with a genome only 1/10 that of humans) provide an opportunity to identify and sequence genes with much less labor.
Conserved synteny can give us an estimate of the number of breakpoints needed to transform 1 map into another. This is a minimum estimate, since internal rearrangements, such as those described in the pig, will be detected only when gene order is known. An even more dramatic representation of such conserved synteny can be obtained directly by chromosome painting, which can also detect interruptions of synteny within a chromosome.
Gaps in the Maps
The numerous gaps in the comparative map are very obvious. Some valuable genomes are poorly represented and include, surprisingly, nonhuman primates and the commercially important horse and dog.
Relatively few primate species have been analyzed thoroughly, despite early interest in primate gene mapping by somatic cell genetics. There has been little ongoing interest in somatic cell genetic mapping, and the suite of loci mapped in different species is idiosyncratic and difficult to compare. One difficulty has been the lack of accurate family data to conduct linkage analysis, a problem now being corrected with systematic studies of multigenerational pedigrees in baboons (Rogers and VandeBerg 1998).
The paucity of comparative data for important domestic species like horse and dog may be surprising but applies more to the choice of markers than the absence of a map. Up to the time of this writing, the emphasis in both species has been on type I markers, and relatively few coding genes are located whose homologues are obvious in other species. Inclusion of more coding genes will enable the rapidly developing linkage maps in these species to be aligned to the human and other maps, to the mutual benefit of all species.
Not so surprisingly, more exotic species like shrew and vole are poorly represented (presumably because we do not eat them and because they are not commercially important) and unfortunately the edentates--thought to have diverged earliest in the eutherian radiation, are completely missing (won't someone construct an armadillo or a sloth gene map?). Farther afield, there have been sustained efforts to construct good maps of commercial bird and fish species, but hardly any data have been reported about reptiles, except for a few assignments in the alligator.
Comparative Painting
Chromosome painting patterns are very much easier to compare between species than are the G-band homologies that were the subject of enormous effort in the 1980s (Rofe and Hayman 1985; Yunis and Prakash 1982). Although detailed G-banding comparisons can potentially provide more information than comparative painting (such as about orientation within conserved segments), in reality, 1 band looks much like another in isolation, and it is possible to detect conserved banding patterns only in relatively large regions. Classic banding studies have therefore been useful in comparisons of closely related, but not distantly related, mammals.
The results from comparative painting between humans and other mammals have been assembled for the first time and are presented, with reference back to the human genome, as the poster in this issue entitled "Comparative Chromosome Painting" (Glas and others 1998). Very large blocks of conserved synteny are obvious throughout. Comparative chromosome painting has largely confirmed the conclusions of comparative mapping--that the genome has been very conserved, at least between eutherian orders that diverged about 60 MYA.
The extent of conservation within orders of eutherian mammals is striking. For instance, 20 primate species have been comparatively painted (Wienberg and Stanyon 1998). Among the great apes, only 1 major rearrangement exists--a fusion between 2 acrocentric chromosomes with homology to human chromosome 2. The fused region has been pinpointed by using probes specific to regions of human chromosome 2. Some of the lesser apes, most notably the gibbon, however, show multiple rearrangements (up to 40) compared with humans.
Particularly striking are whole chromosomes that have survived intact, not only within primates but also in other orders. For example, of the 23 human chromosomes, 16 are represented by a single cat chromosome and the remaining 7 by 2, mostly in uninterrupted blocks (O'Brien and others 1997), confirming and extending the results from the comparative gene map. Conservation of entire chromosomes may cover other orders; for instance, human chromosomes 4 and 17 appear intact in carnivores (cat and seal), artiodactyls (pig and cattle), cetaceans (dolphin), and even insectivores (shrew). The X chromosome is entirely conserved within eutherians, as predicted by comparative mapping, although autosomal additions form XY1Y2 systems in some species.
The minimum numbers of autosomal rearrangements between human and other species can be assessed readily from comparative painting patterns. A comparison of the numbers of conserved autosomal blocks between human and other species reveals 23 between human and chimp (that is, only 1 rearrangement within the 22 human autosomes) The human and cat genome share 32 conserved blocks (that is, 10 rearrangements) including 3 internal rearrangements, and other carnivores about the same number (mink 34 and harbor seal 31). Artiodactyls show somewhat more; for instance bovine and human genomes share 56 autosomal blocks (that is, 34 rearrangements). Remarkably, the common shrew, an insectivore and probably as distantly related to primates as any eutherian mammal, shares only 33 conserved blocks.
Painting between rodent species such as mouse and rat reveals strong homologies and a minimum of 33 conserved segments, implying at least 13 rearrangements (Ferguson-Smith and others 1998). However, relationships between the rodent and the human genome are more complex, and definitive comparisons are not yet available, although Ferguson-Smith and others (1998) report that comparative painting reveals 116 conserved blocks shared between mouse and human.
Preliminary evidence from recent cross-species painting in marsupials supports the hypothesis that the marsupial genome is even more conserved than the eutherian genome. Marsupials have a few very large chromosomes, and painting has dramatically confirmed their close relationships across species (Toder and others 1998).
PRACTICAL VALUE OF COMPARATIVE GENOMICS
Since the mammalian genome is very conserved, it makes sense to combine the gene mapping information from humans and nonhuman mammals for the mutual benefit of both. Because excellent resources are available to the Human Genome Project, both in funding and in the great variety of phenotypes studied worldwide, the detail and quality of information about the human genome is unrivaled, probably forever. Applying our knowledge of the human genome to the genomes of other mammals, birds, and fish of interest as sources of food, clothing, transport, or companionship, or as valued environmental resources, will be of immediate value in improving and conserving these resources. The availability of a detailed human map helps to establish an outline map of any other mammalian species to which details may be added and from which comparative information may be gathered. The resulting information is extremely valuable for animal health and animal breeding.
Genetic information can flow in the opposite direction equally well. Mapping information in a nonhuman species can be directly transferred to the human gene map by reference to conserved chromosome regions. Many studies impossible in humans can be set up in model mammals, which have many advantages for genetic study--including the possibility of genetic manipulation.
Applications to Animal Health
Information from the Human Genome Project can be applied directly to veterinary research, diagnosis, and treatment. The location and identification of disease genes in humans can immediately assist in identification of the homologous condition in animals. If a similar phenotype maps within a syntenic segment conserved between humans and a domestic animal species, it is likely that a mutation in the same gene causes the condition. This knowledge can be applied directly to diagnosing the condition in animals. If treatment regimes are already available for the homologous disease in humans, they may be readily transferred to veterinary use. If the human gene has already been cloned, it is easy to clone the homologous gene from any other animal, and to screen it for mutations. Ultimately, the biochemical cause of the disease may be understood and new treatment or a cure found.
Inherited diseases in dogs and horses are of particular concern. Many dog breeds were developed from very small founder populations subject to selection for different attributes of appearance or behavior and may be highly inbred. Inherited eye diseases are particularly common. Some breeds of horses have been selected over 100 yr for performance on the racetrack, without much regard to health status. A number of diseases (such as bone diseases) are extremely common (Bailey and Binns 1998). Many diseases that are well studied in humans (such as, anemias, muscular dystrophies, immune deficiencies, cancer, and heart disease) occur in domestic animal species, and the knowledge we have gained of the human conditions could have important benefits in animal husbandry.
One good example of the interaction of genetic studies in human and livestock species is the identification and cloning of the gene that controls malignant hypothermia, a condition economically important in pigs and medically important in humans. The discovery of a genetic factor that affected anesthetic response (malignant hyperthermia) in humans was followed by its identification as an economically important stress syndrome in pigs and its location on the pig genetic map. The conserved syntenic region in humans was searched for candidate genes, resulting in the cloning of the human ryanodine receptor gene RYR1 and subsequently, of the pig homologue (Fujii and others 1991).
In addition to providing a direct comparison with human diseases, comparative maps make possible the comparison of disease states between livestock species or between livestock species and model mammals like the mouse (with, for example, kidney diseases).
Animal Breeding
A detailed gene map of an animal species can assist animal breeding, either indirectly by marker-assisted selection or directly by cloning the genes responsible for phenotypes of interest or concern--genetic diseases that are a problem or economic traits that we seek to improve. Crosses are first set up to establish genetic linkage of the phenotype to polymorphic markers on the map. Placement on the bovine map has included several diseases as well as the polled gene, which controls horn development (Womack 1998). Also mapped is the potentially useful callipyge gene, which produces muscular hypertrophy in sheep hindquarters, and 2 genes that affect fecundity (Broad and others 1998). When linkage is discovered, closely linked markers can be used to predict the phenotype of offspring, since the allelic combinations will tend to be parental. Using marker-assisted selection, superior animals may be chosen for raising and breeding at an early stage (even before birth), and animals that will develop a disease may be excluded from breeding.
Although marker-assisted selection has been of some value, cloning the gene that confers the trait of interest is the ultimate goal. A gene map may be used to isolate an unknown gene physically by a technique called positional cloning. One constructs a detailed linkage map and locates the phenotype (desirable or undesirable) with respect to flanking DNA markers by making crosses between animals that differ in this trait. It is then necessary to switch reference to a physical map, constructed by ordering cloned sequences of DNA, to determine the position of the DNA markers flanking the disease gene on a physical piece of DNA (usually a large insert clone like a YAC or cosmid, or a contig of overlapping pieces). The piece(s) of DNA containing the flanking markers must also contain the unknown gene, and they can then be physically searched for sequences with the hallmarks of genes, or sequences transcribed into RNA in the tissue of interest. Candidate genes may be identified and sequenced in normal and variant animals to identify the one revealing base alterations that correlate to the mutant phenotype.
The positional cloning procedure may be short circuited by using a comparative candidate positional cloning approach. This shorter procedure entails locating the trait on a linkage map in the species of interest and then scanning the map of the syntenic region in a species with a high density map (human or mouse) for genes that could be involved. For instance, the fibroblast growth factor gene FGF1 was spotted in a region of the human map syntenic to a region of the sheep map containing DNA markers linked to wool fiber diameter differences. FGF1 was subsequently identified as a major determinant of this important economic trait (Broad and others 1998).
When a gene responsible for a disease or a superior trait has been isolated, its base sequence is determined and translated (on paper or computer) into the amino acid sequence of its protein product. The amino acid sequence will give us some idea of the normal function of the gene, and the changes in the variant animal will indicate what has gone wrong (or right) in the mutant. Animal breeders will thus have tools to diagnose a variant and to select for or against it. One example is the bovine leukocyte adhesion deficiency gene, which was first mapped and then positionally cloned and identified by mutation analysis (Womack 1998), enabling a direct and easy PCR diagnosis of this economically important trait. Other examples, including the callipyge gene (which may significantly improve meat yield in sheep) and the polled gene (which may be used to select for or against animals with horns), are now the subjects of intense searches in cattle and sheep. Understanding the biochemical action of a cloned gene offers the possibilities of additional improvement by selection or genetic manipulation.
Of particular significance for breeding domestic animals is the capacity to locate, and subsequently isolate, genes that have effects on economically important traits (economic trait loci) like weight, meat quality, wool fiber, fertility, and fecundity. It was traditionally expected that practically all economically important traits are affected by many genes, each with a relatively small effect (so-called quantitative trait loci, or QTLs1). Identifying QTLs has been virtually impossible by traditional breeding techniques because the differences in phenotypes may be slight and often obliterated by environmental effects. An entire branch of genetic analysis (quantitative genetics) has developed to handle such data and use the information in breeding practice. It is now evident that QTLs can be mapped and genes with major effects can ultimately be identified and cloned just like any other gene. For instance, linkage has been established for genes with major effects on growth rate and fatness in pigs (Anderssen and others 1994), as well as several QTLs for milk production in the cow (Womack 1998). Information may be transferred from I species to another, such as instances in which QTLs affecting milk yield have been identified in syntenic regions in sheep and cattle.
In addition to QTLs that confer an economic advantage, a number of abnormal characteristics are determined by QTLs and have been identified in dogs to include heart disease, dislocated hip, narcolepsy, and atopy. These can be mapped in dogs and then the syntenic regions of the human and mouse genome can be scanned for candidate genes.
Practical Benefits to Studies of Human Disease and Development
Genetic information from domestic mammals, model mammals, or even other vertebrates can be of great value for studies of the human genome and genetic disease. Information on maps or markers, and on genes involved with diseases or phenotypes of interest, may be available in a nonhuman species, and can be transferred directly to the human gene map by reference to their position on conserved chromosome regions. Many studies impossible in humans can be set up in model mammals. Nonhuman mammals present many advantages for genetic study, including shorter generation time, large family size, and above all the ability to set up crosses deliberately to study the transmission of genes over 3 or more generations. To extend these advantages, it is now possible to genetically manipulate model mammals by targeted disruption of a gene ("knockout") or insertion of a foreign gene (transgenesis).
Many disease states or traits with homologues in humans are well studied in domestic mammals. More than 350 inherited recessive disorders are known in different dog breeds that are particularly accessible to study because of the availability of large multigeneration pedigrees and because within a breed, a particular mutant allele is certain to be identical by descent (Binns and others 1998). For example, eye diseases common in several dog breeds have human homologues that can be studied much more easily in a dog model. Similarly, sheep may provide models for cystic fibrosis, hemophilia, Batten-Mayou disease, congenital cataract, and inherited deafness (Broad and others 1998). It is sometimes advantageous to study a model for a human genetic disease in an animal of a similar size and with similar physiology such as sheep or pig. Many human diseases have homologues in 1 or more animal species, and a detailed compendium (Mendelian Inheritance in Animals) has been developed to assist this comparison (Nicholas and Harper 1996).
It has been important to develop disease models in rodents. More than 1000 spontaneous mutants in mice cause diseases, and 100 of these have been proposed to be homologues of known human genetic diseases (Davisson and others 1998). For example, mutants responsible for inherited deafness have homologues in mutants in several mouse genes, as do several classic genetic defects such as Waardenburg syndrome and Hirschsprung's disease (CGOW 1996).
It is even possible to manipulate the genome in model mammals to create mutants in a candidate gene and study their effects on phenotype. Recessive diseases can be studied by removing a candidate gene in mouse, either by selection for a mutant embryonal stem cell line (such as producing an HPRT-deficient "Lesch Nyhans" mouse) or the targeted disruption of the homologue of a human gene (such as the CFTR gene to create a cystic fibrosis homologue). Dominant mutant alleles can be studied by introducing the mutant into a mouse; for example, a Huntington's disease model was constructed by transfecting mice with a mutant human HD allele having an expanded triplet repeat (Bates and others, 1997). It is noteworthy that the knockout mutant does not necessarily show a phenotype identical to that of the human disease state.
Studies in model mammals are particularly valuable lot identifying quantitative traits. QTLs are difficult to identify in humans but can be identified by crosses involving model mammals or livestock. Several genes involved in diabetes, obesity, epilepsy, and cancer susceptibility were first identified by crosses in rodents and then the information was transferred to human. As another example, crosses between 2 inbred strains differing in their susceptibility to colon cancer resulted in the identification and mapping of 5 colon tumor susceptibility genes (Davisson and others 1998). This knowledge will assist the identification of candidate genes in syntenic regions in human families with inherited colon cancers. Genes involved in fat distribution and metabolism have been mapped on the pig genome (Anderssen and others 1994). Likewise, behaviors exhibited by particular breeds of dog can be followed by cross-breeding and may result in some insight into human behaviors (Binns and others 1998).
Nonmammalian vertebrates have special practical uses. Some species of oviparous vertebrates are particularly suitable for analysis of development because it takes place in an easily accessible egg, rather than in an implanted mass deep inside a heavily fortified uterus. The chicken has been an important model for decades, and with the transparent zebra-fish egg and embryo, it is easy to identify and study developmental mutants just by observing the living embryo within the first few days after fertilization. Mutagenesis screening has identified thousands of recessive developmental mutants that exhibit specific defects such as dorsoventral patterning of the primary embryonic axes, brain, cardiovascular, and organ development, and these may be recovered and studied (lngham 1997), including even those with severe phenotypes that would not survive even early pregnancy in human or mouse.
DEDUCING THE ANCESTRAL MAMMALIAN GENOME
One of the most exciting uses for comparative data is the investigation of genome evolution in vertebrates. Wittingly or unwittingly, all comparative geneticists--whether they are working on sheep improvement or developmental mutants in zebrafish--are contributing to our knowledge about how the vertebrate genome evolved. This knowledge has great significance for our understanding of ourselves and our animal relatives and tot providing many practical benefits that result from a deeper understanding of normal and abnormal gene structure, organization, and function in mammals, including humans.
Comparing genomes across evolutionary intervals provides another dimension to the Human Genome Project. Extrapolated backwards, such comparisons allow us to deduce the form of the genome of common ancestors: of primates, carnivores--and their common ancestor 60 MYA; or of eutherians, marsupials--and their common therian ancestor 130 MYA; of birds and reptiles--and their common ancestor 300 MYA; and ultimately of fish, reptiles, and mammals--and their common vertebrate ancestor 400 MYA. We can chart genome rearrangements that have occurred to separate lineages, and even the deeper events such as the genome duplications that have occurred at least twice in vertebrate evolution.
Vertebrate Phylogeny
To interpret comparative genomic data, it is necessary to refer to a framework of relationships and approximate divergence dates provided by fossil evidence (Colbert and Morales 1991). A greatly simplified representation of the relationships between major groups discussed here is presented in Figure 1. Increasingly, molecular phylogenies are being constructed from masses of DNA sequence data, although discrepancies between different data sets are still the subject of much debate.
The 3 major groups of extant mammals include 2 infraclasses--Eutheria (placental mammals), and Metatheria (marsupials), diverged about 130 MYA--and the subclass Theria, which contains those diverged from subclass Prototheria (the egg-laying monotremes) about 170 MYA. More than 3750 species of placental mammals in 16 extant orders are distributed widely throughout the world. These orders radiated rapidly from an insectivore-like ancestor in the Cretaceous 60 to 80 MYA, so it is difficult to determine the sequence of their divergence (Figure 1). There are about 250 species of marsupials in 16 families in 3 orders (although higher order taxonomy is the subject of perennial dispute). Marsupials are concentrated in Australasia (16 families), with a significant presence in South America (3 families) and 1 North American species. Australian and South American marsupials diverged about 80 MYA. All monotremes are confined to Australasia. There are only 3 species in 2 families, which diverged 30 to 70 MYA. Fossil evidence, as well as their anatomy and physiology, has traditionally placed them as a separate subclass of Mammalia, which diverged independently from the therian (eutherian-marsupial) line of descent about 170 MYA. However, it is possible that monotremes are more closely related to marsupials than eutherians, and this idea received some support from mitochondrial DNA sequence (Janke and others 1995).
Mammals diverged from a branch of reptiles (synapsids), leaving no other descendants. They are therefore equally distantly related to the other 2 major branches of reptiles, diapsids (snakes, crocodiles, and the ancestors of birds) and anapsids (turtles), all of which diverged independently 300 to 350 MYA. Reptiles in turn diverged from amphibians, which evolved from a branch of the fish 350 to 400 MYA.
Establishing Ancestral Gene Arrangements
Based on the foregoing descriptions of comparative genomic data, it is thus possible to make comparisons at vastly different evolutionary levels. Comparing the human genome with those of the great apes, or the mouse genome with rat, can inform us of recent changes within a lineage. Comparing the primate genome with that of carnivores or artiodactyls provides information about changes over approximately 60 MYA, and comparisons between any of these groups and rodents informs us of changes occurring in the last 80 million yr. We can double the evolutionary distance by comparing eutherian mammals with marsupials and treble it by looking at monotremes. Then we can double it again by making comparisons between mammals and birds, reptiles, or fish.
By lining up the syntenic associations between members of a species group, it should be possible to identify similarities and deduce the karyotype of their common ancestor. The data are now available to begin this task, at least for the common ancestors of primates, carnivores, and ungulates, and perhaps even for the ancestral eutherian.
Karyotype arrangements shared by 2 species either could be ancient and retained by both (shared ancestral, or plesiomorphic) or could be the result of a recent change within that particular lineage (shared derived, or synapomorphic). Deducing ancestral arrangements of conserved chromosome regions depends on distinguishing ancestral and derived arrangements. This must be done with reference to an outgroup, a species from a more distantly related group than that to which the study species belong. For instance, cat or pig would provide a suitable outgroup for comparisons between different primates, whereas monotremes or chicken might be a suitable outgroup for comparisons between marsupials and eutherians.
For chromosomes that show rearrangements between human and another species, it is usually possible to deduce the point at which the rearrangement occurred by comparison with a third group. For example, human chromosome 2 is represented by 2 acrocentric chromosomes in chimpanzee and gorilla. Which state is ancestral? Did the change represent a fusion in the human lineage or a fission in a common chimpanzee-gorilla lineage? Clearly the question is important, not only for deducing how human chromosome 2 came into being, but also in establishing the relationships among the 3 great ape species. The answer is clear because all higher primates have the 2 acrocentrics, which must therefore be the ancestral condition. Indeed, close examination of the sequences around the region reveals a relic of an abandoned centromere close to the fusion point (Wienberg and Stanyon 1998). This scenario is also consistent with other data, suggesting that humans are more closely related to chimpanzee than either is to the gorilla.
A good example of the application of these principles at a deeper evolutionary level is the investigation of mammalian sex chromosome evolution (Graves 1995). Comparative mapping shows that X chromosomes of eutherians are almost invariant. However, human X-linked genes fall into 2 very distinct groups in marsupials: Genes on human Xq and the pericentric region are on the X in all marsupials, but markers distal on the short arm are autosomal. This could mean either that the ancestral X was large like the eutherian X and lost a portion to autosomes in marsupials or, conversely, that the ancestral X was small like the marsupial X and gained autosomal regions in eutherians. Appeal to a third group of mammals, the more distantly related monotremes, favors the latter possibility, since human Xp markers map to similar autosomal clusters in the platypus.
The Big Picture
How near are we to describing the ancestral eutherian, marsupial, and monotreme genomes? An ancestral mammalian genome? An ancestral reptilian or even vertebrate genome?
From the results of comparative gene mapping and comparative painting, it is possible to trace large sections of the genome that are the same in closely and even distantly related eutherians. An ancestral primate genome, using shared synteny between human, Old World, and New World monkeys, could be reconstructed, allowing identification of the lineage-specific rearrangements such as the multiple rearrangements in the gibbon genome (Wienberg and Stanyon 1998).
Curiously, however, more rearrangements separate different primates than separate humans from some nonprimate mammals. For example, 17 rearrangements exist between human and lemur, but only 7 between human and cat, not counting internal rearrangements. This immediately tells us that many of the changes occurred in the lemur lineage and that cat and human more closely represent the ancestral genome. This comparison is the first step in reconstructing a putative primate-carnivore ancestral genome (Rettenberger and others 1995). Confirmation of this high degree of conservation between the 2 orders comes from comparisons between human and other carnivores; mink with 8 rearrangements and seal with 10.
These comparisons may be extended to other orders. More rearrangements (from 14 to 22) were described between humans and ungulates, at first suggesting a more distant relationship. However, the finding of only 7 rearrangements between humans and dolphin, a member of the whale family now thought to have diverged more recently from ungulates, implies that many of the differences observed in syntenic associations and painting patterns between human and sheep, pig, cattle, muntjac, and horse genomes are likely to be specific for the artiodactyl and/or perissodactyl lineage.
Adding rodent genomes into the equation complicates the analysis immediately since more than 90 rearrangements separate human from mouse and rat. Does this reflect simply an increased divergence date'? Comparison with more distantly related groups (outgroups) suggests that many of these rearrangements occurred in the rodent lineage. For example, human chromosomes 21 + 3 appear as a unit in every eutherian group except rodents and also appear to be intact in marsupials. Ten genes on human chromosome 2 are split between 2 mouse chromosomes, but all lie in the same linkage group in zebrafish. In agreement with this conclusion, chromosome painting between human and the common shrew (an insectivore, thought to be the most distantly related eutherian) identified only 33 conserved blocks, implying that primate and insectivore genomes differ by only 10 rearrangements.
It should therefore be easiest to deduce the form of the genome of a common eutherian ancestor by comparing common chromosome blocks between the most conserved of the distantly related species: human, cat and/or seal, dolphin, and shrew. Inspection of the comparative painting poster (Glas and others 1998) points to several autosomes that appear intact in each of these species (as well as in some or all of the ungulates): These include conserved regions represented by human chromosomes 3, 6, 9, 11, 13, 17, 18, and 20. In addition, a number of human autosome regions are associated in all other species, suggesting that they were ancestral but disrupted in the primate lineage. For example, associations of human 3/21, 14/15, and 16/19 are present in cat, bovine/pig, dolphin, and shrew. In this way, it should be possible to build a picture of the genome of an ancestral therian that lived 80 to 130 MYA.
Deducing an ancestral genome in the other 2 branches of mammals has been less of a challenge. Marsupials have a few large chromosomes that made possible some of the most thorough classical studies of karyotype evolution in any mammal group. Extraordinary karyotypic conservation has enabled an ancestral marsupial karyotype to be deduced by cytological criteria alone, even before the discovery of G-banding, certainly before comparative gene mapping had much impact, and long before the advent of chromosome painting. A "basic" 2n--14 karyotype, with nearly identical G-band patterns, is represented within each of the major marsupial groups, and other marsupial karyotypes are easily derived from it (Rofe and Hayman 1985). Even in the karyotypically diverse kangaroo family, karyotypes can be related by Robertsonian fusions and fissions. The 2n= 14 basic karyotype itself may have been derived from a 2n--22 ancestral marsupial karyotype (Svartman and Vianna-Morgante 1998).
Comparative gene mapping of marsupial autosomes is not yet sufficiently advanced to test the hypothesis put forward by Role and Hayman, although at least the groupings of autosomal genes appear to be the same in kangaroos and dasyurids, which diverged about 45 MYA. Chromosome paints have been prepared from marsupial species (Ferguson-Smith and others 1998) and painting is being adapted rapidly to marsupial chromosomes. At the time of this writing, the published results largely confirm (with some interesting exceptions) the predictions of Role and Hayman for kangaroo karyotypes (Toder and others 1998).
Monotremes have a few large and many small chromosomes, and the 3 extant species have karyotypes almost G-band identical (Graves 1998). The involvement of several small unpaired chromosomes in a translocation chain at meiosis (a feature unique among mammals) is also shared among the 3 species, although the numbers of chromosomes involved is different between the platypus and the 2 echidna species. Comparative mapping has confirmed the same gene arrangements on the X and 2 of the largest autosomes in platypus and echidna.
Attempts to compare the maps of eutherians, marsupials, and monotremes to deduce an ancestral mammalian karyotype are somewhat premature, given the paucity of autosomal markers on the maps. As yet it has not been possible to paint autosomes across such vast evolutionary distances, although cross-species painting of the human X by the wallaby X has been achieved (R. Glas, La Trobe University, Victoria, Australia, personal communication, 1998).
Genome Stability
The limited karyotypic change in marsupial evolution has usually been regarded as an oddity of a weird group of mammals, and the conservation of the monotreme karyotype, an artifact of the paucity of species. However, comparative mapping and painting now present a picture of an extremely stable mammalian genome in which rapid change is the exception. Indeed, the conservation of synteny between human and bird and fish maps suggests most strongly that the vertebrate genome is extremely stable. It is the variability of some eutherian karyotypes--especially rodent--that is out of line.
Different eutherian groups show very different degrees of genome stability. Primates show karyotypic similarities identifiable by G-banding (as well as comparative mapping) and beautifully confirmed by chromosome painting between species that diverged as long as 50 MYA. However, the gibbon is exceptional, showing a dramatically scrambled gene map as well as a rainbow of colored stripes on chromosome painting (Wienberg and Stanyon 1998). Among the carnivores, the cat family has an almost invariant karyotype, but the dog family reveals much variation. In the same way, different marsupial groups evidence different levels of variation, from the dasyurids with almost no karyotypic variation among many species, to the macropodids with a spread of haploid numbers and chromosome morphologies. At the extreme are the rock wallabies, in which more than 20 different karyotypes are found in a very rapidly diverging species complex (Toder and others 1998).
What is it that makes a karyotype stable? Is it in some way an intrinsically good genome arrangement with some sort of selective advantage (what?)? Or does something happen to destabilize the genome in 1 lineage'? At the time of this writing, we still do not understand the role of genome change in speciation. Recent work suggests that interspecific hybridization could play a role in rapid genome remodeling by unleashing bursts of transposon activity (O' Neill and others 1998).
Genome Duplication in Vertebrates
Can comparative gene mapping help us track the major changes that occurred in the shaping of the vertebrate genome more than 400 MYA? It was suggested nearly 3 decades ago that genome duplication took place during the early diversification of vertebrates (Ohno 1970). The confirmation since then of groups of paralogous genes in mammals has been interpreted as evidence for 2 doublings of the vertebrate genome. However, it has been difficult to distinguish polypoidization from tandem duplications of localized regions. Mapping 144 loci in the zebrafish has revealed very large duplicated segments in both species, suggesting that 2 polypoidization events occurred before the divergence of the mammal lineage from fish 450 MYA (Postlethwaite and others 1998).
Evolution of Genetic Control Mechanisms
Comparative genomics can inform us not only about the evolution of genome arrangements but also about how different genetic control mechanisms evolved and how they function. Comparative studies of recombination, genomic imprinting, X-chromosome inactivation, and sex determination provide good examples of the advantages of comparative studies.
Linkage mapping depends on recombination, which occurs during meiosis in the male and female parent. Analysis of the products of meiosis in the 2 sexes in humans and other eutherians reveals a minor deficit in recombination in males. However, this deficit does not appear to represent a general rule of the influence of sex on recombination since in marsupials, females have far less recombination than males as the result of a strongly sex-dependent distribution of chiasmata (Bennett and others 1986). This major variation in chromosome behavior during meiosis may help clarity the molecular basis of initiation of recombination.
Knowing the physical location of imprinted genes and studying them in more than 1 mammal species are critical to the study of genomic imprinting. Genomic imprinting refers to the expression of an allele according to its parental origin (Tilghman 1992). For instance, the IGFII gene on the mouse chromosome 7 is expressed only from the paternal copy (de Chiara and others 1991); however, a nearby gene, H19, is expressed only from its maternal copy. Intense mapping around these loci in mouse has uncovered several other imprinted genes in the region, suggesting that parental imprinting is under regional control. The finding that these genes also map together on human chromosome 11p, and are also imprinted, implies that imprinting evolved more than 70 MYA, and is therefore likely to be important for survival or reproduction. Several human genetic diseases are caused by accidental derivation of both homologous chromosomes from the same parent (uniparental disomy), mutation of imprinted genes, or disruption of genomic imprinting. Mapping of the same genes in marsupials and studying their expression will enable conserved elements of the molecular mechanism (methylation?) to be identified. These investigations will also help us to determine when, and perhaps why, imprinting evolved. Is it a dosage compensation mechanism? Or is it a protection against parthenogenesis? Or does it reveal an "arms race" between the male and female genomes to commandeer resources for the embryo?
X-chromosome inactivation is a large-scale control mechanism affecting the activity of thousands of physically linked genes on 1 X chromosome in females. Discovered in 1961, it is still unclear at the time of this writing how X inactivation works. Mapping genes to the X in human and mouse and studying their expression has been critical to progress in understanding how this system of transcriptional cis control--over megabases of DNA--is exerted by the XIST locus on the X (Rastan 1994). Mapping the same genes in other eutherian mammals, and particularly in marsupials, has been important in understanding the molecular changes that accompany inactivation, since marsupials show an incomplete, less stable inactivation and paternally imprinted form that is probably ancestral (Cooper and others 1993). Since there appear to be differences in the molecular mechanism between eutherian and marsupial X inactivation, comparisons will help to identify common elements (methylation? histone deacetylation?) common to the 2 systems. It will also be critical in understanding why X inactivation evolved: Was it simply a dosage compensation mechanism to ensure fair play between XX females and XY males? Or was it a primitive dose-dependent sex-determining mechanism?
Sex determination has been an area of intense study in mammals, beginning with the search for the testis-determining factor on the Y chromosome. Comparative gene mapping and cloning, especially in human, mouse, and marsupial, has played interacting roles in the search for testis-determining factor and the investigation of how it, and other genes with which it interacts, fulfills its role as trigger of the male-determining pathway (Pask and Graves 1998).
CONCLUSION
Comparative genomics has advanced significantly since the last comparative genomics report (Comparative Genomics Organization Workshop 1996). At the time of this writing, genetic and physical maps are being constructed for a variety of species using a variety of methods, and comparative chromosome painting has been applied very rapidly to cytogenetic comparisons of homology.
Genetic mapping is well advanced in several livestock species as well as rodent model species, and outline maps are available for at least 30 mammal species in 8 orders. Maps are being constructed rapidly for chicken and fish species. Although the level of detail between maps of different species may differ by 2 orders of magnitude, it begins to be possible to compare the genomes at a greater or lesser level of resolution. Arrangements may then be compared across mammal--and even other vertebrate--species.
Comparative mapping enables transfer of information between human, livestock, and rodent genomes. Comparisons of location can be used to identify homologous genes involved in disease states in humans, domestic animals, and rodent models, making possible the development of techniques for diagnosis and treatment for transfer in either direction. Pinpointing a genetic trait on a linkage map is the first step to physically isolating it; and when it is isolated from 1 species, it may be obtained from any other by homologous cloning. This enables us to identify genes that control genetic diseases, and even makes it realistic to consider cloning genes involved in quantitative traits (economic traits like weight or milk yield, as well as diseases like cardiovascular disease or cancer). Comparative gene mapping can therefore deliver information for the benefit of research into animal health and animal breeding. It should greatly speed up the search for genes that specify inherited diseases in mammals and humans, as well as genes that specify economically important traits.
Comparisons, even over vast evolutionary time scales, show that the mammal genome--indeed, the vertebrate genome--has been very conserved. Thus it now becomes possible to ask how the mammal genome--even the vertebrate genome--evolved, and how it works.
1Abbreviations used in this paper: BAC, bacterial artificial chromosome; FISH, fluorescence in situ hybridization; G-banding, Giemsa banding; MYA. million years ago; PCR, polymerase chain reaction; QTL, quantitative trait locus; RFLP, restriction fragment length polymorphism; YAC, yeast artificial chromosome.
REFERENCES
Anderssen L, Haley CS, Ellergren H, Knott SA, Johansson M, Andersson K, Andersson-Eklund L, Edfors-Lilja l, Fredholm M, Hansson I, Hakansson J, Lundstrom K. 1994. Genetic mapping of quantitative trait loci for growth and fatness in pigs. Science 263:1771-1774.
Bailey E, Binns MM. 1998. The horse gene map. ILAR J 39:171-176.
Bates GP, Mangiarini L, Mahal A, Davies SW. 1997. Transgenic models of Huntington's disease. Hum Mol Genet 6:1633-1637.
Bennett JH, Hayman DL, Hope RM. 1986. Novel sex differences in linkage values and meiotic chromosome behaviour in a marsupial. Nature 323:59 60
Binns M, Holmes N, Breen M. t998. The dog gene map. ILAR J 39:177-181.
Broad TE, Hill DF, Maddox JF, Montgomery GW, Nicholas FW. 1998. The sheep gene map. ILAR J 39:160-170.
Burr DW, Cheng HH. 1998. The chicken gene map. ILAR J 39:229-236. Colbert EH, Morales M. 1991. Evolution of the Vertebrates. New York: Wiley-Liss.
CGOW [Comparative Genome Organization First International Workshop J. 1996. Comparative genomic organization of vertebrates. Mamm Genome 7:717-734.
Cooper DW, Johnston PG, Watson JM, Graves JAM. 1993. X-inactivation in marsupials and monotremes. Semin Dev Biol 4:117-128.
Davisson MT, Bradt DW, Merriam JJ, Rockwood SF, Eppig JT. 1998. The mouse gene map. ILAR J 39:96-131.
de Chiara TM, Robertson EJ, Efstradiatis A. 1991. Parental imprinting of the mouse insulin-like growth factor 2 gene. Cell 64:849-859.
Elgar G, Clark M. 1998. The pufferfish gene map. ILAR J. 39:249-256.
Ferguson-Smith MA, Yang F, O'Brien PCM. 1998. Comparative mapping using chromosome sorting and painting. ILAR J 39:68-76.
Fujii J, Otsu K, Zorzat F, de Leon S, Khanna VK, Weiles JE, O'Brien PJ, MacLennan DH. 1991. Identification of a mutation porcine ryanodine receptor associated with malignant hyperthermia. Science 253:448-451.
Glas R, Wakefield M J, Toder R, Graves JAM. 1998. Comparative chromosome painting. (Poster). ILAR J 39.
Graves JAM. 1995. The origin and function of the mammalian Y chromosome and Y-borne genes--An evolving understanding. BioEssays 17:311-320.
Graves JAM. 1998. Gene maps of monotremes (Mammalian subclass Prototheria). ILAR J 39:225-228.
lngham PW. 1997. Zebrafish genetics and its implications for understanding vertebrate development. Hum Mol Genet 6:1755-1760.
Janke A, Gemmell NJ, Feldmaier-Fuchs G, von Haeseler A, Paabo S. 1995. The mitochondrial genome of a monotreme--The platypus (Ornithorhynchus anatinus). J Mol Evol 42:153-159.
Levan G, Stahl F, Klinga-Levan K, Szpirer J, Szpirer C. 1998. The rat gene map. ILAR J 39:132-137.
Lyons LA, Laughlin TF, Copeland NG, Jenkins NA, Womack JE, O'Brien SJ. 1997. Comparative anchor tagged sequences (CATS) for integrative mapping of mammalian genomes. Nat Genet 15:47-56.
Morizot DC, Nairn RS, Walter RB, Kazianis S. 1998. The linkage map of Xiphophorus fishes. ILAR J 39:237-248.
Nadeau JH, Taylor BA. 1984. Lengths of chromosomal segments conserved since divergence of man and mouse. Proc Natl Acad Sci U S A 81:814-818.
Nesterova TB, lsaenko AA, Matveeva NM, Shilov AG, Rubtsov NB, Vorobieva NV, Rubtsova NV, VandeBerg JL, Zakian SM. 1997. Novel strategies for eutherian x marsupial somatic cell hybrids: Mapping of the genome of Monodelphis domestica. Cytogenet Cell Genet 76:115-122.
Nesterova TB, Mazurok NA, Rubtsova NV, lsaenko AA, Zakian SM. 1998. The vole gene map. ILAR J 39:138-144.
Nicholas FW, Harper PAW. 1996. Inherited disorders: The comparative picture. Aust Vet J 73:64-66.
O'Brien SJ, Seuanez HN, Womack JE. 1988. Mammalian genome organization: An evolutionary view. Ann Rev Genet 22:323-351.
O'Brien S J, Wienberg J, Lyons LA. 1997. Comparative genomics: Lessons from cats. Trends Genet 13:393-399.
Ohno S. 1967. Sex Chromosomes and Sex-linked Genes. Heidelberg: Springer-Verlag.
Ohno S. 1970 Evolution by Gene Duplication. Heidelberg: Springer-Verlag.
O'Neill RJW, O'Neill M J, Graves JAM. 1998. Undermethylation associated with retroelement activation and chromosome remodelling in an interspecific mammalian hybrid. Nature 393:68-72.
Pask A, Graves, JAM. 1998. Sex chromosomes and sex determining genes--Comparative analysis in mammals and other vertebrates. Genetica (Forthcoming).
Postlethwaite JH, Johnson SL, Midson CN, Talbot WS, Gates M, Ballinger EW, Africa D, Andrews R, Carl T, Eisen JS, Home S, Kimmel CB, Hutchinson T, Johnson M, Rodriguez A. 1994. A genetic map for the zebrafish. Science 264:699 703.
Postlethwaite JH, Yan Y-L, Gates MA, Home S, Amores A, Brownlie A, Donovan A, Egan ES, Force A, Gong A, Goutel C, Fritz A, Kelsh R, Knapik E, Liao E, Paw B, Ransom D, Singer A, Thomson M, Abdul-jabbar TS, Yelick P, Beier D, Joly J-S, Larhammar D, Rosa F, Wester field M., Zon LI, Johnson SL, Talbot WS. 1998. Vertebrate genome evolution and the zebrafish gene map. Nat Genet 18:345-349.
Rastan S. 1994. X-chromosome inactivation and the Xist gene. Curr Opin Genet Dev 4:292-297.
Rettenberger G, Klett C, Zechner U, Bruch J, Just W, Vogel W, Hameister H. 1995. ZOO-FISH analysis: Cat and human karyotypes closely resemble the putative ancestral mammalian karyotype. Chromosome Res 3:479-486.
Role R, Hayman D. 1985. G-banding evidence for a conserved complement in the Marsupialia. Cytogenet Cell Genet 39:40-50.
Rogers J, VandeBerg JL. 1998. Gene maps of nonhuman primates. ILAR J 39:145-152.
Rubstsov NB. 1998. The fox gene map. ILAR J 39:182-188.
Samollow PB, Graves JAM. 1998. Gene maps of marsupials. ILAR J 39:203-224.
Serov OL. 1998. The American mink gene map. ILAR J 39:189-194.
Serov OL, Matyakhina LD, Borodin PM, Searle JB. 1998. The common shrew gene map. ILAR J 39:195-202.
Svartman M, Vianna-Morgante AM. 1998. Karyotypical evolution of marsupials: From higher to lower diploid numbers. Cytogenet Cell Genet (Forthcoming).
Tilghman SM 1992. Parental imprinting in the mouse. In: The Harvey Society Lectures, series 87. The Harvey Society of New York. New York: Wiley-Liss. p 69-84.
Toder R, O'Neill RJW, Graves, JAM. 1998. Chromosome painting in marsupials. ILAR J 39:92-95.
Venta PJ, Brouillete JA,Yuzbasiyan-Gurkan V, Brewer GJ. 1996. Gene-specific universal mammalian sequence-tagged sites: Application to the canine genome. Biochem Genet 34:321-341.
Wakefield MJ. 1998. Internet comparative mapping resources. ILAR J 39:66-67.
Wakefield MJ, Graves JAM. 1998. Comparative genome maps of vertebrates. (Poster). ILAR J 39.
Wienberg J, Stanyon R. 1998. Comparative chromosome painting of primate genomes. ILAR J 39:77 91. Womack JE. 1998. The cattle gene map. ILAR J 39:153-159.
Yunis JJ, Prakash O. 1982. The origin of man: A chromosomal pictorial legacy. Science 215:1525-1530.
FIGURE 1 Relationships of the major groups of vertebrates. Fossil evidence reveals a very approximate time scale for the divergence of reptiles from fish and amphibians, the divergence of mammals from synapsid reptiles, the divergence of the 3 major mammal groups, and the divergence of the eutherian orders. AMPH, amphibians; NW, New World; OW, Old World.
Copyright © 2008. National Academy of Sciences.
All rights reserved.
500 Fifth St. N.W., Washington, D.C. 20001.
Terms of Use and Privacy Statement