Michael Ashburner
Cambridge University,working in collaboration with the Berkeley Drosophila Genome Project (Gerry Rubin, Sima Misra, Suzanna Lewis and colleagues)
The BDGP have virtually completed the sequence of the "Adh region" - that is the 69 polytene chromosome bands from 34C4 to 36A2. The sequence is in two large contigs with one gap (of unknown size, but not visible by in situ with flanking probes) and a single unattached P1 clone at the distal end. The sizes of these are 1,901,227-bp and 762,658-bp and 82,692-bp, a total of 2,746,577-bp.
Conventional genetic analysis identified 65 genes in this region by virtue of their mutant alleles. In addition, a further 12 genes are predicted because of the phenotypes of deletion overlaps. Of the 65 known genes, 48 have at least one lethal allele, and for seven sterility is the severest known phenotype. 31 of these genes have P-element insertion alleles. Over 850 chromosome aberrations in this region have been characterised.
Our objective is both to predict all of the genes in this chromosome region from the sequence and to identify all of the genetically known genes at the sequence level. Using a variety of methods - ab initio algorithmic gene predictions, sequence matches, EST hits, etc, 200 genes are predicted. Of these, 188 are protein coding, 11 are tRNAs and one is a probable snRNA pseudogene. Of the 77 genetically known (or predicted) genes, 51 have been identified at the sequence level, and we have reasonable candidates for most of the others (which are now being tested). 78 of the 188 genes have Drosophila EST hits; a further 14 are hit by a Drosophila EST from a related gene. Over half (102) of the protein coding regions have significant BLAST matches with sequences from mouse or human, 65 with C. elegans (of which 59 also have mouse &/or human matches), 42 with S. cerevisiae and 22 with Arabidopsis.
The average gene density, one gene per 13.7kb, is less than on the X tip (see EDGP abstract); the region includes 18 transposable elements (one per 15.2kb, similar to that on the X tip). There are at least seven examples of "genes within genes" and several examples of gene-pairs, that is, closely related sequences in tandem. By screening collections of P-elements (especially those from the EP screens) eight genes known only from gene predictions have P-element insertions. These are now being systematically jumped to determine whether or not these genes are required for phenotypic normality.
No large scale analysis of the sequence has yet been done. One very obvious question is the relationship between the sequence and the polytene chromosome banding map; this will be studied. One obvious question is how do these data inform estimates of total gene number in Drosophila? What a simple question! Simple arithmetic (and assuming a "euchromatic" genome size of 115-mb) says 7,000. But this may be misleading if the gene density is abnormally low in this region (the X tip sequence data indicate 14,500 genes). 69 polytene chromosome bands are 1.34% of the total: using this as a basis for calculation estimates 13,500 genes. Although these numbers should be treated with some caution they all indicate that the fly may have fewer genes than C. elegans (18,500, Steve Jones, personal communication).
From genome mapping to disease genes
Grant R. Sutherland
Department of Cytogenetics and Molecular Genetics, Women's and Children's Hospital, Adelaide
Much mapping of the human genome has taken place over the last decade, there are now many different maps available. These include high resolution linkage maps, STS maps, YAC contig maps, EST maps, radiation hybrid maps and integrated maps containing information from all of these. The next generation of maps will be the SNP (single nucleotide polymorphism) maps which are being rapidly constructed now. Within a few years the ultimate map, the annotated sequence will be completed. Examples of the use of such maps for the location and identification of human disease genes will be given. These will include primarily genes on human chromosome 16, however other examples including epilepsy genes will be presented. The process of going from map position to gene identification, positional cloning, has undergone considerable changes over the last five years as more detailed maps of the genome have been constructed. Currently, many projects of disease gene identification rely on a positional candidate gene approach, rather than positional cloning.
A scan for linkage disequilibrium across the human genome
Gavin Huttley
Human Molecular Genetics, John Curtin School of Medical Research, ANU, Canberra 0200
Linkage disequilibrium (LD), the tendency for alleles of linked loci to co-occur nonrandomly on chromosomal haplotypes, is an increasingly useful phenomenon for: 1) revealing historic perturbation of populations including founder effects, admixture or incomplete selective sweeps; 2) estimating elapsed time since such events based on time-dependent decay of LD; and 3) disease and phenotype mapping, particularly for traits not amenable to traditional pedigree analysis. Because few descriptions of LD for most regions of the human genome exist, we searched the human genome for the amount and cM extent of LD among 5,048 autosomal STRP (short tandem repeat polymorphism) loci ascertained as specific haplotypes in the Caucasian CEPH mapping families. Approximately 4% of STRP loci separated by less than 4.0cM are in LD and the fraction of locus pairs within these intervals that display LD is directly proportional to the inverse of recombination distance between them (1/theta). A 5cM sliding window analysis across 22 autosomes revealed 9 genomic segments (0.9-6.2cM), where multilocus LD is elevated above that presented by the HLA on chromosome 6, a region known to have extensive LD. These regions, including HLA on chromosome 6, are likely indicators of historic selective sweeps that highlight genome segments of human evolutionary consequence.
Experimental evaluation of microsatellite-based estimates of effective population size and inbreeding in bottlenecked populations
P.R. England, R.A. Nichols, D.A. Briscoe and R. Frankham
School of Biological Sciences, Macquarie University, NSW 2109
When a population passes through a bottleneck it experiences inbreeding and loses genetic diversity due to random genetic drift. These processes threaten bottlenecked populations with extinction by lowering their evolutionary potential and fitness. Genetic drift and inbreeding occur in proportion to the effective population size (Ne). This idealised measure is usually very different to the apparent size of a population, especially in populations that have recently recovered from a bottleneck. Therefore, reliable estimators of Ne and inbreeding are necessary if vulnerable, wild populations are to be identified and appropriately managed. Microsatellites promise to lend unprecedented precision to estimates of Ne and inbreeding in natural populations. We evaluated the reliability of microsatellite-based estimates of Ne and inbreeding applied to experimentally bottlenecked populations of Drosophila. Two genetic methods for estimating Ne were evaluated: a method that measures changes due to drift in allele frequencies over time and a method that measures linkage disequilibrium in a single population sample. Both methods provided estimates that correlated significantly with theoretical Ne but which had low precision. The temporal change method produced overestimates and the linkage disequilibrium method produced underestimates. Two methods of estimating F were also evaluated: a maximum likelihood method (FMLE) that is based on changes due to drift over time in allele frequencies and an estimate based on loss of heterozygosity (FDH) over time. Precision of FMLE was higher than FDH. The direction in which estimates diverged from expectation depended on the severity of the bottlenecks experienced.
Molecular ecology of fur seals
Neil Gemmell
University of Canterbury, Private Bag 4800, Christchurch, New Zealand
Despite intensive study, the population structure and breeding system of fur seals remain unresolved. Behavioural observations are difficult given the sheer density and number of animals involved, while mark-release-recapture and radio telemetry studies require large amounts of time and effort to return useful data. An alternative approach is to use molecular genetic techniques to investigate relationships within and between fur seal populations.
Microsatellites are widely regarded as the best genetic markers yet discovered. Thirty-five pinniped microsatellite primer sets have been tested on a variety of phocid and otariid species. Fifteen of these amplify discrete polymorphic loci in fur seals. These markers are now being used to produce a detailed picture of social organization and population substructure in the South American fur seal (Arctocephalus australis) and the Antarctic fur seal (Arctocephalus gazella). Initial analyses of male reproductive success and population structure in these two species are presented.
Patterns of genetic variation in two species of Australian freshwater fishes
Dugald J. McGlashan and Jane M. Hughes
Co-operative Research Centre for Tropical Rainforest Ecology and Management,
Australian School of Environmental Studies, Griffith University, Queensland
The comparative approach can be useful when attempting to distinguish between the importance of dispersal, isolation, distance and earth history on levels of population genetic structuring within a species. Pseudomugil signifer (pacific blue-eye) and Craterocephalus stercusmuscarum (fly specked hardyhead) are small bodied freshwater fish with qualitatively similar life histories. They are sympatrically distributed in lowland north eastern Australian drainages, although C. stercusmuscarum occurs in south eastern and northern drainages, and above waterfalls. Fish were collected from the Northern Territory to south east Queensland (sympatrically where possible) and allozyme techniques were used to estimate population genetic structure. Both species had high levels of population structuring with overall FST estimates of approximately 0.6. However, patterns inherent in the data are strikingly different. Among P. signifer populations, most of the variation is distributed among "regions" followed by among drainages and lastly within drainages. C. stercusmuscarum, in contrast, was comprised of two distinct groups: the Northern Territory and upland populations of east flowing rivers forming one group, and lowland populations of east flowing rivers forming the second. Results will be discussed in terms of the interaction of intrinsic and extrinsic factors that contribute to the evident mosaic population genetic structures, and the taxonomic status of some Australian freshwater fishes.
Dugong population genetics as inferred from control region sequences
D. Tikel, D. Blair and H.D. Marsh
University of Queensland
Intraspecific genetic differentiation of the dugong was inferred using hypervariable region 1 (HVR-1) sequences of mitochondrial DNA (mtDNA). HVR-1 was sequenced for 103 dugongs from Australia, West Indian Ocean and Asia. Among these sequences, 39 variable sites and 37 haplotypes (specific DNA sequences) were indicated. Phylogenetic trees constructed from the haplotypes showed three clusters: West Australian, East Australian and Asian. These haplotype clusters suggest that Asian and East Australian dugongs are more closely related to each other than either group is to the West Australian dugongs. Surprisingly, haplotypes of West Indian Ocean dugongs are extremely similar to the haplotypes of dugongs from East Australia despite their large geographical separation. The geographical range of the Asian mtDNA haplotypes does not overlap with the Australian haplotypes. The geographic ranges of the two Australian mitochondrial DNA genetic units overlap in the region of the Great Barrier Reef. We propose that the differentiation of these two genetic units may be attributed to Torres Strait acting as a periodic land bridge, preventing dugong movements during the Pleistocene low sea level phases. Using a series of AMOVA (nested analyses), substructure is indicated among dugong populations of the Australian coastline in a manner that is suggestive of isolation-by-distance.
Determinants of rate variation in mammalian DNA sequence evolution
Lindell Bromham
Department of Zoology, University of Queensland
Attempts to analyse variation in the rates of molecular evolution among mammalian lineages have been hampered by paucity of data and by non-independent comparisons. I used phylogenetically independent comparisons among mammal species to test several explanations for rate variation which predict correlations between rate variation and generation time, metabolic rate and body size. I compared the relative rates of sequence evolution for mitochondrial and nuclear genes, protein-coding, rRNA and non-translated sequences from 61 mammal species representing 14 orders. Correlation analyses performed on differences in genetic distance since common origin of each pair against differences in body mass, generation time and metabolic rate reveal that substitution rate at four-fold degenerate sites in two out of three protein sequences is negatively correlated with generation time. In addition, there is a relationship between the rate of molecular evolution and body size for two nuclear-encoded sequences. No evidence is found for an effect of metabolic rate on rate of sequence evolution. This pattern - that rate of substitution varies with body size/life history across mammalian species - suggests a number of possible causes of variation in substitution rate between lineages.
Population viscosity and multi-nest colonies in the ant Polyrhachis doddi
R.N. Johnson and R.H. Crozier
La Trobe University
The presence of two or more fertile queens in a single social insect colony has been termed ëThe second socialityí and demands explanation in terms of population genetics and ecology. As polygyny does not immediately fit classical kin selection theory, Polyrhachis doddi, a facultatively polygynous weaver ant, is being used to analyse the evolutionary impact of such behaviours on social structure. Where multiple queens occur together the degree to which they are related is not known. It is also uncertain whether P. doddi has a polydomous (multi-nest) colony structure. Nests often occur in clusters on a single tree, but the genealogical relationships among them are unclear. We used microsatellite analysis to compare relationships both among individuals from within nests, between nests within clusters and among individuals between clusters.
Intra-colony relatedness estimates suggest that individuals are more highly related (0.2 < r < 0.5) than expected for a polygynous and polydomous species, and that mated daughters are probably recruited back into the natal nest. Pairwise G-tests showed non-significant differences in population genotype distribution between nests from the same tree. This result suggests that nests within clusters could be regarded as single polydomous colonies, but gene flow appears to be restricted between nests from different trees, even between clusters from a single population reflecting high population viscosity (FST = 0.2922).
Dodging extinction and breaking the rule: the generation of genetic variation in parthenogenetic aphids
Alex Wilson*, Paul Sunnucks and Dinah Hales
School of Biological Sciences, Macquarie University, NSW 2109
*Current address: Division of Botany and Zoology, Australian National University, ACT 0200
Genetic variation is important to the long-term survival of an organism and parthenogens are assumed to have limited means with which to generate such variation. The absence of chromosomal reassortment and meiotic recombination between homologous chromosomes in apomictically parthenogenetic aphids would seem to be of great disadvantage to their long-term survival. However, we have evidence of long-term parthenogenetic reproduction and potentially important ecological differentiation amongst closely related Sitobion clones.
Our program addresses the levels of genetic variation in parthenogenetic and sexually reproducing aphids and also explores possible cryptic mechanisms for the generation of genetic variation in parthenogenetic aphids. Previous karyological studies have suggested the possibility of a mini-meiotic event occurring during the apomictic production of male embryos. An additional cryptic mechanism for the generation of genetic variation during parthenogenesis is mitotic recombination. Recently, we have been using X-linked and autosomal microsatellite loci to address these phenomena in addition to our on-going study of the evolution and population biology of Australian and New Zealand Sitobion.
Morphological variation and allelic distribution across a hybrid zone between two forms of Geleznowia verrucosa (Rutaceae)
Linda Broadhurst1, David Coates2 and Beng Tan3
1School of Environmental Biology, Curtin University of Technology, GPO Box U1987, Perth, Western Australia 6845
2Department of Conservation and Land Management, Perth, Western Australia
3School of Environmental Biology, Curtin University of Technology
Geleznowia verrucosa Turzc. (Rutaceae) is a monotypic but morphologically diverse genus, endemic to the sandplains north and east of Perth Western Australia. Three forms (small, large and intermediate) can be distinguished according to leaf and flower dimensions and flower abundance. Populations are generally small and disjunct, with forms rarely co-occurring. Morphological and genetic data strongly suggest the small and large forms represent distinct taxa with the intermediate form being of hybrid origin. Allozymes and morphometrics were used to examine a rare population of co-occurring small and large forms and a zone of integration between them. Morphometrically, the small and large forms appeared to maintain their integrity. Individuals in the integration zone, which could not be readily characterised according to form, exhibited affinities with both the small and large forms. Allelic distribution, particularly in the five aspartate aminotransferase (Aat) loci, indicates asymmetric gene flow from the large form to the small, is occurring. This data further supports the hypothesis that the small and large forms are systematically distinct, with the intermediate form being of hybrid origin.
DNA markers, QTL mapping and the theory of junctions
Ian R. Franklin
CSIRO Animal Production, Blacktown, NSW 2148
Mathematical expressions have been obtained for the expected lengths of chromosome segments of different parental origin in intercross or backcross populations segregating for an arbitrary number of genetic markers per chromosome. This theory has a number of implications for gene mapping, for detecting QTL and for other uses of gene marker technology. For example, tests for the presence of QTL are usually based on comparing the observed effect of a marked chromosomal region against a null hypothesis of no effect on the trait. If the two parental lines from which the populations were derived differ substantially for the trait under investigation, as they usually do, this is invalid because each marked region carries with it a substantial region of chromosome from one or other parent.
The theory discussed here allows the expectations under the null hypothesis to be calculated correctly. It also describes the expectations under an infinitesimal model, and examples are shown of apparent large QTL effects in well designed experiments where there are, in fact, no loci of large effect on the trait present. It also lays the groundwork for an approach to mapping that focuses on defining the positions of recombination points (junctions) rather than marker positions.
Some of the implications of this theory for other applications of DNA marker technology, such as introgression or grading up, are also discussed.
Precise mapping of QTLs in intercrosses between highly divergent breeds of pigs using multiple trait analysis
Gerhard Moser
Department of Animal Science, University of Sydney, NSW
Department of Animal Husbandry and Breeding, Division of Animal Breeding, University of Hohenheim, Stuttgart, Germany
Usually, experimental designs for mapping QTL include simultaneous measurements of many individual traits. The mapping of QTL for each trait, separately, normally increases the number of relevant significant QTL involved. The resolution of mapping QTL by marker analysis is a major factor affecting practical application from QTL experiments. Much attention has focused on improving the efficiency of marker analysis by taking into account the dependence of quantitative traits on several linked or unlinked loci. However, single trait analysis can not differentiate whether the effects are due to a single locus with correlated effects on several traits or are due to two or more loci each affecting a single trait. Apart from an increase of the statistical power and more precise parameter estimation, joint mapping allows testing for pleiotropy versus close linkage as a cause of genetic correlations. Especially in pigs, where genetic response is based on the improvement of several negatively correlated traits, breaking of unfavorable linkage offers new breeding opportunities. A computer program for the analysis of multiple traits in crosses between outbred populations has been written and properties of the analysis are illustrated on simulated and real data.
Disease gene mapping in dogs
Alan Wilton
School of Biochemistry and Molecular Genetics, University of NSW, Sydney, NSW 2052
Most dog breeds have a recessive genetic disease that is common. This is a consequence of the breeding structure of high use of champion animals and breeding to close relatives. The loci responsible can be located by gene mapping. This could result in elimination of the disease from the breed through identification of carriers and selective breeding. The dog genome map is in its infancy with less than 200 loci assigned. We are using microsatellites in the border collie to map two genes, collie eye anomaly (CEA) and ceroid lipofuscinosis (CL). CEA is a congenital eye defect with variable expression. CL is a nerve degenerative disease similar to Batten disease in humans. At about 18 months affected animals develop symptoms of behavioural changes with gradual loss of control of mental and motor nerves.
We have typed 30 loci in border collie pedigrees with a small number of affecteds. Lod score analysis has not shown close linkage of the CL gene to any these loci. Work on a similar disease in English setters by collaborator Frode Lingaas in Norway has identified linked markers using test crosses. Even though one marker is within 4cM of the disease gene we have not been able to confirm linkage to CL in the border collie. It is possible that the diseases are caused by different disease genes in the two breeds. In humans there are 5 loci known that cause CL-like disease. Three have been identified and these are being used as candidate loci in the dog. Conservation of synteny is being assumed to identify whether one of the other two loci are the cause of dog CL. Loci known to be close to the unidentified human genes are used to identify polymorphisms in the dog that can be used in a mapping study to identify linkage.
Mapping genes for manic depression (bipolar affective disorder)
J.A. Donald1, L.J. Adams2,3, P.B. Mitchell3 , S.L. Fielder2, A. Rosso2 and P.R. Schofield2
1School of Biological Sciences, Macquarie University, Sydney
2Garvan Institute of Medical Research, Sydney
3School of Psychiatry, University of New South Wales & Mood Disorders Unit, Prince Henry Hospital, Sydney
Bipolar affective disorder is characterised by severe mood swings (mania and depression) and affects 1-2% of the population. Despite evidence that there is a strong genetic component to manic depression, the genetic basis is not understood. Many groups are currently undertaking genome screens and several putative susceptibility loci have been identified, although replication of such findings has proved problematical. Possible reasons for this will be discussed. We have recently completed a genome screen using 214 microsatellite markers on 35 individuals from the most powerful pedigree in our Australian cohort. The data were analysed by two-point linkage analysis under several diagnostic models. Four markers gave results possibly indicative of linkage. A more intensive screen was undertaken in the regions surrounding these markers using 52 additional members of this family. Results for markers on chromosome 4q35 remained positive and meet the criteria for "suggestive" linkage. Analysis of a further ten pedigrees indicated this predisposing locus may be present in at least one additional family. These results suggest the presence of a novel bipolar disorder susceptibility locus on chromosome 4q35. Approaches for confirming and identifying such a locus will be discussed.
Molecular evolution of circadian clock genes
Bambos Kyriacou
Department of Genetics, University of Leicester, Leicester, UK
A number of clock genes that are central to circadian pacemaking have been identified in insects, fungi and mammals. The two best understood molecules are encoded by the period (per) and timeless (tim) genes, and their primary role in the Drosophila clock has been elucidated. Both are involved in a mutual negative feedback loop which generates relentless cycling of their mRNAs and, after a 4-6h lag period, rhythms in their products. The lag between transcript and protein, and the degradation kinetics of the products are two crucial parameters for generating an oscillation. The negative feedback model, however, may need to be extended or revised, because other forms of per regulation are present, at least in Lepidoptera. We have used Musca domestica as a comparative model and observe patterns of PER regulation that are presently difficult to reconcile with a simple negative feedback model. The use of various chimeric genes between different species per genes, have also revealed unexpected features of PER regulation, and can be used as a powerful tool to study the interactions of these clock molecules in the Diptera. Finally, chimeric clock genes have produced some fascinating insights into the control and evolution of species-specific behaviours.
Involvement of rad21 in sister chromatid cohesion and DNA double strand break repair
William D. Warren, Matthew Beasley, Diane Severin and Michael McKay
DNA Repair Laboratory, Peter MacCallum Cancer Institute, Melbourne
DNA double-stranded breaks (dsb) are one of the most deleterious genetic lesions with which a cell must cope. How effectively these lesions are repaired significantly contributes to the fate of the cell after DNA damage. The isolation and genetic analysis of ionizing radiation-sensitive mutants of bacteria and yeast have contributed substantially to our understanding of the cellular mechanisms involved in the repair of radiation-induced DNA damage. In eukaryotes, DNA dsb can be repaired via two main pathways: (1) recombinational repair, which is dependent on the presence of an intact homologous DNA sequence, and (2) end-to-end rejoining, which involves non-templated ligation of the two DNA ends. In mammals, DNA dsb appear to be repaired primarily by end-to-end rejoining whereas in lower eukaryotes dsb repair occurs almost exclusively by the recombinational pathway.
The Schizosaccharomyces pombe rad21 mutant is severely affected in its ability to reconstitute chromosomes broken by gamma-irradiation, strongly suggesting the rad21 gene product is required for rejoining gamma-irradiation induced DNA dsb. Here we report the cloning and characterization of human, mouse and Drosophila homologues of rad21 and present preliminary data exploring the involvement of these and related proteins in dsb repair, sister chromatid cohesion and meiotic recombination. Transgenic Drosophila ectopically expressing rad21-GFP fusions have recently been made and confocal microscopic analysis of protein localization and expression in living embryos is under way. Alterations in the human rad21 gene are currently being examined in DNAs from cancer patients that exhibit acute sensitivity to radiation therapy. Progress towards the isolation of Drosophila rad21 mutants will be also be presented as will progress towards the generation of a knockout mouse.
Epigenetic effects on transgene expression in mice can be inherited through meiosis
Heidi Sutherland, David Garrick, Margot Kearns, David Martin* and Emma Whitelaw
Department of Biochemistry, University of Sydney, NSW 2006
*Fred Hutchinson Cancer Research Center, Seattle, USA
It is now clear that gene expression in complex multicellular organisms is affected by epigenetic processes i.e. processes whereby genes are rendered inactive which do not involve changes in the DNA sequence itself. These modifications play an integral role in X-inactivation and genomic imprinting in mammals and it is likely that epigenetic modifications are crucial for developmentally-controlled expression at many endogenous genetic loci. This type of gene silencing is also observed in yeast, Drosophila and plants.
Epigenetic modifications at transgenes often result in mosaic expression. This variegated or heterocellular expression has largely been overlooked because it can only be detected if transgene expression is assayed at a single-cell level. In our studies we have used a system in which the E. coli lacZ gene is driven by transcriptional control elements derived from the human globin genes. Staining of erythrocytes for b-galactosidase expression permits convenient assessment of the proportion of erythrocytes that express the transgene. We have spent the last few years characterising the factors governing this transgene silencing. These factors include the cis-acting elements e.g. enhancers present in the transgene, the number of copies of the transgene in the integrated array, the integration site and the age of the mouse.
Epigenetic modifications are usually considered to be somatic and erased on passage through the germline, ensuring the renewal of totipotency. At some loci in plants, yeast and Drosophila, epigenetic modifications have been found to persist through meiosis. We have a transgenic mouse line in which the epigenetic modification appears to be inherited through meiosis. The implications of this finding will be discussed.
The involvement of the patched/hedgehog signalling pathway in genetic disease, common human cancer and developmental anomalies
Brandon Wainwright
Centre for Molecular and Cellular Biology, The University of Queensland, St Lucia, Qld 4072
The hedgehog/patched signalling pathway is conserved in development from flies through to mammals and has been implicated in tumourigenesis by its involvement in the cancer predisposition syndrome nevoid basal cell carcinoma syndrome (NBCCS; Gorlin's syndrome), as well as in a range of sporadic tumours. Patched encodes a transmembrane protein which acts as the receptor for the secreted glycoprotein hedgehog in a heteromeric receptor complex involving a third molecule, smoothened (smo). In Drosophila patched functionally opposes the action of hedgehog such that the activation of downstream target genes is a balance between patched activity and hedgehog signalling from adjacent cells. In mammalian cells the identities of the downstream genes have yet to be determined, although based upon studies in Drosophila one might predict that at the least target genes will be members of the Wnt gene family and the TGF-b superfamily, including the bone morphogenic proteins (BMPs).
As a result of a positional cloning experiment we demonstrated that the human homologue of patched is defective in patients with NBCCS, a syndrome where patients present with a predisposition to basal cell carcinoma, medulloblastoma and a variety of developmental defects. Subsequently we were able to demonstrate that patched is mutated in the common, sporadic form of basal cell carcinoma and medulloblastoma. Most of those mutations are truncations and they are distributed throughout the patched molecule and we have also identified several missense mutations. Genotype/phenotype analysis has not revealed any association between the position of the truncation and the presence/absence of major and minor features of the disease. However, the four missense mutations we have found so far appear non-random in their distribution. Intriguingly, these mutations are commonly found in the patched transmembrane domains. Recent work on the disease Niemann-Pick type C (NPC) has indicated that the the NPC protein shares homology with patched and that this homology is present in the transmembrane domains, thought in the case of NPC to be sterol sensing domains. This raises the possibility that patched responds to sterol, possibly when complexed to its ligand sonic hedgehog. In addition to the role of patched in BCCs and medulloblastomas, mutations in the human smoothened gene (SMO) have also been shown to produce sporadic BCCs as well as squamous cell carcinomas (SCCs).
In order to examine the function of patched we have modelled a number of human mutations in Drosophila.
Do sleepy lizards sleep around? A molecular genetics test of lizard morals
Steven J.B. Cooper1, C. Michael Bull2 and Ben C. Baghurst2
1Evolutionary Biology Unit, South Australian Museum, Adelaide, SA
2School of Biological Sciences, Flinders University, Adelaide, SA
Behavioural studies have shown that the sleepy (stumpy) lizard, Tiliqua rugosa, is socially monogamous with pair fidelity over more than one mating season, representing one of the only reported cases of such a behaviour in reptiles. The nature of this social monogamy involves no parental care of offspring, and provides an interesting variation of the form of social monogamy practised by many bird species. Social monogamy does not necessarily imply sexual monogamy and genetic analyses have revealed extra-pair fertilisations in many apparently monogamous species. In this paper we report the results of paternity tests of sleepy lizard families using microsatellite markers, to determine whether social monogamy equates to sexual monogamy in this species. The genetic data are compared with behavioural observations of these lizards carried out on a 70 ha site near Mt. Mary, South Australia, over five breeding seasons. Of 21 litters examined, with a total of 42 offspring, 17 (81%) showed that the female's partner had fathered all the offspring, and 4 (19%) showed evidence of extra pair fertilisation (EPF). Although sample sizes were small, females of polygynous males were more likely to experience EPF. The behavioural and genetic data are consistent with the hypothesis that social monogamy in a season results from male sleepy lizards guarding their female partners.
Ribosomal DNA ITS sequences help to resolve relationships among parasite species
J.A.T. Morgan
Department of Parasitology, University of Queensland, Brisbane, Qld 4067
Many soft bodied invertebrates, and especially parasites, offer few morphological features to permit distinguishing among species. However, identification of these parasites is vital for accurate diagnosis of disease, or recognition of pest species. Full identification in some species, for instance trematodes, might only be possible by experimental completion of their complex life cycle. Material of a single stage is often insufficient for specific identification using morphological criteria. Use of genetic methods offers much promise for identifying cryptic species. Genetic variation at the nucleotide level provides the highest resolution available for investigating heritable differences. Ribosomal DNA has proved a useful tool for distinguishing among trematode species within a number of genera. Less work has been carried out on monogenean species. Results from analyses on the trematode genus Echinostoma and the monogenean genus Entobdella indicate that the internal transcribed spacers provide sufficient variation to discriminate among closely related species. Preliminary studies suggest that the phylogenetic resolving power of the ITS2 is limited because alignment is only straightforward among members of the same family. The use of ITS2 secondary structures as a guide to aid sequence alignment above the family level was investigated.
Can fluctuating asymmetry be used to monitor inbreeding and loss of genetic variation in endangered populations?
Dean M. Gilligan, Lynn M. Woodworth, Margaret E. Montgomery, David A. Briscoe and Richard Frankham
Key Centre for Biodiversity and Bioresources, School of Biological Sciences, Macquarie University, NSW 2109
Controversy surrounds the relationship of fluctuating asymmetry (FA), a measure of developmental stability, with loss of heterozygosity and inbreeding. FA analysis has been proposed as a simple technique for identifying inbred or environmentally stressed populations. FA of sternopleural bristle number in Drosophila melanogaster was measured to test for a relationship between FA and both allozyme heterozygosity and pedigreed inbreeding coefficients. FA of populations maintained at effective sizes of 25(8 replicates), 50(6), 100(4), 250(3) and 500(2) for 50 generations (inbreeding coefficients of 0.05 to 0.71) failed to show a significant relationship with allozyme heterozygosity or the pedigreed inbreeding coefficient. A second comparison of highly inbred (F ~ 1) lines and their outbred base population also failed to show any significant relationship between FA and inbreeding. Our results do not support the use of FA as a conservation tool for identifying inbred populations or those suffering a loss of genetic variation. Data is currently being compiled for a meta-analysis and will be presented if available.
Physical mapping of a single copy gene on the centromere of wheat chromosome 1B
Michael Francki1 and William Berzonsky2
1Dept. Botany, University of Western Australia, Nedlands, WA, 6907
2Dept. of Plant Sciences, Loftsgard Hall, North Dakota State University, ND 58105, USA
The centromere is an important structure enabling proper segregation of chromosomes. Although a considerable amount of information is available for mammalian centromeres, little is known about the DNA organization of plant centromeres. Several studies have identified repetitive DNA families within cereal centromeres but the nature of other DNA sequences is largely unknown. RFLP markers have been mapped to the centromeres of wheat chromosomes, however it is difficult to determine whether these markers are on the centromere or on the proximal region of chromosome arms. In this study, we have used a physical mapping approach to identify RFLP markers on the centromere of wheat chromosome 1B. Wheat-rye translocations that contain the short arm of rye chromosome 1 fused to the long arm of wheat chromosome 1B with centromeric breakpoints were used in the analysis. Our results, using the rye-specific centromeric sequence, pAWRC.1, indicate that 1BL.1RS lines contain a small portion of the centromere from 1R. One RFLP marker identified a small 1BS segment in all 1BL.1RS lines. This segment maps proximal to pAWRC.1 in 1BL.1RS and on the centromere of 1B. Characterisation of this cDNA will be presented and its possible role in centromere function is discussed.
Human genetic disease associated microsatellite loci and platypus population study
Shiro Akiyama, Neil J. Gemmell, Neil D. Murray and Jennifer A. Marshall Graves
School of Genetics and Human Variation, La Trobe University
Twelve human genetic diseases have been identified in association with expansion of microsatellite repeats. All loci are highly polymorphic in normal human populations. Some loci, including FMR-1 (Fragile X Mental Retardation) and Huntington's disease, are highly conserved across vertebrates and also polymorphic in other mammals. Therefore, the possibility of the application of human genetic disease associated microsatellite locus as a genetic marker for platypus population study and also as a 'universal' genetic marker for mammals were examined.
Locus FMR-1 (CGG and AGG repeat) was successfully amplified in the platypus with a total of nine alleles and heterozygosity of 55.4%. Surprisingly, 56.6% of heterozygosity was detected in male platypus (n=76) since FMR-1 is localized on X chromosome in humans (Xq27.3) and other eutherian mammals. Also locus PL 12.1 (CCT repeat), which was obtained by screening the platypus genomic DNA library, showed high homology with human Huntington's disease locus (homology 68%). This locus was also polymorphic in the platypus populations with a total of five alleles and 29.1% of heterozygosity in mainland populations. Tasmanian populations appear to contain a common null allele.
Multiple forms of a mitochondrial gene in individual worms
Lynne van Herwerden and David Blair
Dept. Zoology & Tropical Ecology, James Cook University, Townsville, Queensland
Efforts to use partial ND1 sequences to infer a phylogeny for species of the genus Paragonimus (Trematoda: Digenea) were confounded by the discovery of at least two ND1 sequences within individual worms. The divergence of the ND1 types is shown, by phylogenetic analysis, not only to predate the divergence of the three Paragonimus species groups investigated, but also the divergence of some trematode families. Some sequences appear to be pseudogenes, as they have a single base deleted and, when translated, they contain premature termination codons. The implications for phylogenetic studies in trematodes, using ND1 sequence data will be discussed.
Models for genetic control, with application to the common carp (Cyprinus arpio)
S. Davis, E. Catchpole, G. Fulford and R. Pech
School of Mathematics and Statistics, University College, UNSW, ADFA
Several forms of control are currently under consideration for control of carp (Cyprinus carpio) in Australia. Perhaps the most novel is a transgenic method which makes use of either an inducible fatality gene (IFG) or an inducible sterility gene (ISG). The transgene is integrated into the genome of target populations by stocking of transgenic animals. The IFG or ISG is then "triggered" to control population abundance. We develop simple models for the integration of a selectively neutral transgene. We investigate the effectiveness of various management strategies whereby the population is periodically exposed to the trigger substance. We show that there exists a fundamental tension between keeping large proportions of the population transgenic (and therefore susceptible to the trigger) and reducing abundance of breeding stock by killing or sterilising large numbers of transgenic individuals. Good improvements in speed of integration can be gained by releasing young which carry multiple copies of the transgene.
Mitochondrial DNA evidence for rapid colonisation throughout the Indo-west Pacific by the mudcrab Scylla serrata
David Gopurenko1, Jane Hughes1 and Clive Keenan2
1Australian School of Environmental Studies, Griffith University, Nathan , QLD 4111
2Bribie Island Aquaculture Research Centre, PO Box 2066, Bribie Island, QLD 4507
The green mud crab Scylla serrata is widely distributed throughout mangrove habitats of Indo-Pacific coastal waters. It is generally assumed that high levels of population admixture and gene flow occur both within and between mudcrab populations. This assumption is based on the belief that female mudcrabs release their eggs offshore suggesting high levels of mixing of propagules before entry into coastal adult habitats. In order to ascertain levels of historical and contemporary gene flow for this species, this study investigates the phylogeographic distribution of mitochondrial DNA haplotypes sampled throughout the species range. Adults were sampled from three west Indian ocean locations (N=21), five west Pacific sites (N=28) and four sites from northern and eastern Australia (N=98). From temperature gradient gel electrophoresis and sequencing of 549 base pairs of a mtDNA gene (cytochrome oxidase 1) twenty distinct haplotypes were identified. The distribution of these haplotypes indicates limited contemporary gene flow between trans-oceanic sites. Genealogical assessment of sequenced haplotypes suggests that historical spread of S. serrata throughout the Indo-Pacific has occurred rapidly and recently (<1Myr bp) from a west Pacific origin. Results also suggest a long-term vicariant event occurred to isolate north Australian crabs prior to or during the Indo-Pacific radiation, such that the region contains an ancestral 'race' of mud crab.
Natural selection and the human mitochondrial genome
Simon Easteal
Human Genetics Group, The John Curtin School of Medical Research, The Australian National University
The replacement of most early humans by the descendants of a small African group is indicated by several features of human mtDNA variation: relative sequence homogeneity, unimodal Poisson distributions of pairwise differences among sequences in most populations, and the occurrence of deeper branches in Africa than any other location. Other (but by no means all) genetic, archaeological and paleontological evidence has been interpreted as supporting this model; it is the mtDNA, however, that provides the most compelling evidence, but the strength of this evidence depends on an assumption of selective neutrality. The pattern of human mtDNA variation could equally have resulted from the global spread of a selectively favoured mtDNA lineage - a selective sweep. These two alternative possibilities can be evaluated by comparison of the pattern of mtDNA variation with i) neutral expectations, ii) variation of genes in the nuclear genome, and iii) variation in other species, particularly chimpanzees. We have found that the pattern of human mtDNA variation is inconsistent with neutral theory predictions, it is inconsistent with nuclear genome patterns and it is inconsistent with the level of variation in chimpanzees. We conclude that the human pattern of mtDNA variation has arisen as a consequence of a selective sweep through diverse archaic populations with which we have an ancient and unbroken connection.
Evolution of the b-globin genes: rearranging the branches
Rory Hope1, Steven Cooper2, David Wheeler1, Mi-Hye Lee1, Robert Holland3, Andrew Gooley4, Morris Goodman5
1Department of Genetics, University of Adelaide, South Australia
2Evolutionary Biology Unit, South Australian Museum
3School of Physiology and Pharmacology, University of New South Wales
4Macquarie University Centre for Analytical Biotechnology, New South Wales
5Department of Anatomy and Cell Biology, Wayne State University, Detroit, Michigan
The human haemoglobins were amongst the first proteins to be sequenced and, based on sequence comparisons, Ingram (1961) (Nature 189: 704-708) proposed a model to account for the evolution of the globin genes. The model, which Ingram suggested illustrated a general evolutionary phenomenon, incorporated the notion of successive gene duplications and sequence divergences giving rise to increased developmental and functional complexity amongst the members of the resulting gene family. This was the first demonstration of the use of molecular data to derive the evolutionary relationships of paralogous genes. The analysis of extensive globin DNA and protein sequence data from a wide variety of species has provided strong general support for Ingram's original model and has led to the derivation of a gene tree that is probably unsurpassed in the level of detail that it contains. Nevertheless, many questions concerning globin gene evolution remain unanswered. In this presentation we show how the molecular phylogenetic analysis of b-globin genes from marsupials and monotremes can provide, and has provided, crucial information about: i) the order and timing of the b-globin gene duplication events in relation to major species divergences in mammals, and ii) the changes that have taken place in the developmental expression of the b-globin genes. The recent discovery by Holland & Gooley (1997) (Eur. J. Biochem. 248: 864-871) of a b-like globin protein (w-globin) in the Tammar Wallaby (Macropus eugenii) that has sequence affinities to avian b-globin proteins has prompted us to characterise the w-globin gene. Our preliminary results (see poster presentation by Wheeler et al.) suggest the need to incorporate an additional and ancient gene duplication event into the accepted model of b-globin gene evolution. This study provides yet another example of the use of genetic studies of marsupials to uncover variants of commonly accepted evolutionary theories.
Field and laboratory heritabilities: some lessons from Drosophila
Ary A. Hoffmann
Department of Genetics, La Trobe University, Bundoora, Vic 3083
In a widely-cited paper, Roff and Mousseau (Heredity 58: 103-118, 1987) summarized lab estimates of narrow sense heritability for traits in Drosophila falling into four categories (morphology, life history, physiology and behaviour) and concluded that life history and behavioral traits had low heritabilities, whereas morphological and physiological traits had high heritabilities. It has been suggested that similar patterns occur under field conditions. This has been interpreted as evidence that some classes of traits are under more intense selection than others and that traits differ in the nature of environmental inputs. In this talk I will review recent field and lab studies in Drosophila to address four main questions. Firstly, are the lab patterns consistent with recent lab studies that have markedly increased the number of traits available? Secondly, are comparisons across classes confounded by differences in how accurately traits are measured? Thirdly, do field heritabilities match lab heritabilities and do they differ between trait classes? Finally, can the true field heritability of Drosophila be estimated by comparing lab and field generations? Answers to these questions lead to several surprises and a reevaluation of heritable variation across trait classes.
Hierarchical genetic analysis of an onychophoran species
Paul Sunnucks, Natalie Curach, Jordan French, Dave Briscoe and Noel Tait
Macquarie University, School of Biological Sciences
Analysis of genetic variation in the onychophoran (peripatus) Euperipatoides rowelli has produced some unexpected findings. 1/ Subdivision of allozyme species. Microsatellite and mtDNA analysis agree that allozyme analysis greatly underestimates true species diversity. There is little variation within populations, but adjacent populations can carry fixed differences, and patterns for microsatellites are similar. This paradox occurs against a background of high mtDNA diversity. Thus sex differences appear to be important. 2/ Genetic localization is extremely high. Most unusually, females in some populations show significant isolation-by-distance, while males do not. This pattern is not ubiquitous in space. 3/ Females have paired reproductive tracts, each side with a sperm storage organ (spermatheca). Apparently there are discrete mating episodes, and some individuals hold sperm in only a single spermatheca per mating season. This gives rise to an asymmetry between the paired tracts in stage of development of offspring (but not total number). Genetic data indicate that spermathecae sometimes empty, and are refilled by mating. These features give rise to highly partitioned paternity even within an individual female.
P elements and recombination in Drosophila
John Sved, XiuMei Liang, Mark Tanaka and Yasmine Gray
School of Biological Sciences, University of Sydney
P elements lacking either a left end or a right end are inactive in the genome, even in the presence of P transposase. However activity can be restored if a right-end element on one chromosome is combined with a left-end element at the same site of a homologous chromosome. The result implies that the two ends are capable of finding each other even if they are not physically on the same element. Analysis of the recombinant products reveals that the ends 'excise' as in a normal element, and that two processes, both of which can lead to recombination, are involved in the resolution of the excision event: (1) the two ends remain associated as a 'hybrid element' and insert somewhere nearby in the chromosome (Hybrid Element Insertion - HEI), (2) the ends which do not contain P elements are joined through repair and ligation (Hybrid Excision and Repair - HER).
Engels and Preston introduced the use of very closely linked RFLPs to investigate the mechanism of recombination and the role of DNA repair associated with a single P element. These markers were found by surveying a range of wild type strains, from which a set was chosen which differed maximally from the original P element insertion. To date, RFLPs have not been available for end-deleted elements. We have now incorporated the same set of markers into crosses involving the end-deleted elements, to investigate in more detail whether all classes of recombination can be explained in terms of the HEI and HER processes.
Stabilizing selection in multilocus models
Reinhard Bürger1 and Alexander Gimelfarb2
1University of Vienna
2University of Maryland
Stabilizing selection on quantitative traits eliminates extreme phenotypes and preserves those near the optimum. Therefore, stabilizing selection is considered to exhaust genetic variation. This view has been substantiated by classical analyses that were based on the assumption that loci are independent and allelic contributions infinitesimal. Further support came from Wright's study of the so-called quadratic-optimum model in which two diallelic loci control the character. More recent analyses of two-locus models, however, found that high levels of polymorphism and genetic variance can be maintained by stabilizing selection if allelic contributions at the two loci are not equivalent or symmetric, or if linkage is admitted. I shall report results from mathematical and numerical analyses of 2-, 3-, 4-, and 5-locus systems about the equilibrium structure under stabilizing selection, in particular, about the number of stable monomorphic and polymorphic equilibria, about the probability that trajectories with randomly chosen starting points converge to a given type of equilibrium, about the level of linkage disequilibrium, the average effects of alleles involved in polymorphisms, and the genetic variance maintained. It is found that with four loci or more, stable equilibria involving more than one segregating locus become extremely rare, however, many monomorphic or one-locus polymorphic equilibria can coexist. The ratio of the average genetic variance and the squared average allelic effect decreases rapidly, i.e., proportional to 1/n2. Thus, while in the two-locus case it is the major locus that is segregating, with increasing number of loci, only loci with effects much lower than the average effect are segregating. These results demonstrate that stabilizing selection on quantitative traits can maintain high levels of genetic variation only if the number of contributing loci is not much greater than two.
Molecular analysis of gene-for-gene specificities of the alleles of the flax L gene for rust resistance
J.Ellis1, G.Lawrence1, J.Luck1, P.Dodds1, M.Ayliffe1, K.Shepherd2, D.Frost1 and J.Finnegan1
1CSIRO-Plant Industry, GPO Box 1600, Canberra, ACT 2601
2Department of Plant Science, Waite Campus, PMB1, Glen Osmond, SA 5064
Thirteen different rust resistance specificities are genetically mapped to the L locus in flax. The locus contains a single resistance gene with 13 allelic alternatives each encoding a nucleotide binding siteóleucine-rich repeat class (NBS-LRR) protein. Similar proteins in other plant species control resistance to bacteria, fungi viurses and nematodes.
Twelve of the 13 L alleles for rust resistance were cloned and sequenced. An additional L allele (LH) from a variety of flax susceptible to all flax rust isolates was also sequenced, providing a total of 13 alleles for comparison. Ten allelic products were very similar to the product of the originally characterised L6 allele [1]. While variation occurred throughout the amino acid sequences, it was more pronounced in the LRR region, a region favoured for specifictity control. Three alleles had undergone large deletions (L1 and L8) or sequence duplication (L2) in the LRR encoding region of the gene. Transgenic plant experiments using the L2, L6 and L10 alleles indicated that the cloned genes were sufficient for specific rust resistance. Expression of two cDNA clones of L6 using the 35S promoter, indicated that specificity was determined by the coding region and that alternative spliced mRNA products identified in L6 plants that involve the splicing or retention of intron 3 [1], may not be necessary for resistance.
The sequence comparisons and domain swap experiments indicated that the amino terminal region and the LRR region of L proteins both influence specificity. For example, L6 and L11 differ only in the LRR region or the protein product and determine distinct specificities. L6 and L7, however, differ only in the product of exon 1 and are identical in the rest of the protein and also control distinct specificities. The sequence comparisons also indicate that allelic variation is generated by multiple crossing over or gene conversion events that shuffle pre-existing sequence variation into novel combinations. Molecular evolutionary analysis indicates that the LRR-encoding region has undergone diversifying selection.
[1] Lawrence et al. (1995). Plant Cell 7: 1195-1206
Modulation of virulence by the integration of genetic elements in Dichelobacter nodosus
Margaret E. Katz, Gabrielle Whittle and Brian Cheetham
Division of Molecular and Cellular Biology, School of Biological Sciences, University of New England, Armidale, NSW 2351
The Gram-negative anaerobe Dichelobacter nodosus is the causative agent of footrot in sheep. We have characterised three genetic elements, the intA (vap), intB and intC elements, in the genome of D. nodosus. Each element contains an integrase gene. The intA element is present in all virulent strains that we have examined and is absent from most benign strains. In the virulent strain A198 one copy of the intA element is integrated into the tRNA-serGCU gene, close to the askA gene, and a second copy is integrated into the tRNA-serGGA gene, next to the pnpA gene. The intC element is found in most D. nodosus strains, both benign and virulent, and is usually integrated into the tRNA-serGCU gene, next to askA. Between askA and tRNA-serGCU we identified a gene, glpA, whose predicted protein product has very high amino acid identity with RsmA from the plant pathogen Erwinia carotovora. RsmA acts as a global repressor of pathogenicity by repressing the production of extracellular enzymes. We found that in virulent strains of D. nodosus the intA element was integrated next to pnpA, and either the intA or intC element was integrated next to glpA. By contrast, all but one of the benign strains had intB at one or both of these two positions. We showed that the loss of the intC element from the virulent strain 1311 resulted in loss of thermostable protease activity, a virulence factor. We propose a model whereby integration of the intA and intC genetic elements modulates virulence by altering the expression of glpA, pnpA, tRNA-serGCU and tRNA-serGGA.
The dynamics of repeat sequences: application to the epidemiology of Mycobacterium tuberculosis
Mark Tanaka1, Marc Feldman1, Hugh Salamon2 and Peter Small2
1Dept. of Biological Sciences, Stanford University, Stanford, CA 94305, USA
2Division of Infectious Diseases and Geographic Medicine, Stanford University, Stanford, CA 94305, USA
The genome of Mycobacterium tuberculosis harbours a family of insertion sequences (IS6110) that are relatively stable over a short time period (months) but have an observable transposition rate over longer time scales. Using RFLP, this property can be harnessed to generate DNA fingerprints of strains isolated from patients, that in turn allow us to track the spread of TB in a population. One way to summarise the information in these fingerprints is to record the copy number for a given strain of M. tuberculosis (approximated by the number of bands in the relevant fingerprint). A large data-set of fingerprints from TB cases in San Francisco has provided an observed distribution of IS6110 copy number. We propose a stepwise mutation model to describe the dynamics of IS elements in M. tuberculosis, including important aspects of TB epidemiology. Through a comparison of the observations with predictions of the model, we make a qualitative assessment of the various potential forces shaping the distribution of copy number. For instance, it is unlikely that TB in the city of San Francisco is at equilibrium.
Extracting phylogenetic information from noisy data: examples from the arthropods and gastropods
Don Colgan, Peter Eggler, Winston Ponder, Anne McLauchlan, Julie Macaranas and Sue Livingston
Australian Museum, 6 College Street, Sydney 2000
The "first fine careless rapture" about the potential of DNA sequence data for the resolution of phylogenetic questions has become a somewhat distant memory for evolutionary biologists. There are too many examples of discordance for relationships based on molecular data to be automatically preferred to those based on classical morphology. In historical terms, however, molecular phyogenetics is a very young science. It will need a few more decades before investigators have enough information to be confident at the outset of a project that they can make appropriate choices of sequences to be studied. During this period of maturation a large number of sequences will be determined. Questions will arise about the optimal method of using this information, which may be statistically "noisy".
Two large-scale sequencing studies have recently been completed in the Evolutionary Biology Unit. In the first, two regions of 28S rDNA (with consistency indices of 0.556 and 0.652) and histone H3 (0.219) were surveyed in 36 gastropods and four outgroups. In the second, 31 arthropods were sequenced for histone H3 (0.258) and 29 for the U2 snRNA (0.485). Both of these studies are the first to span the morphologically-estimated phylogenetic range of their group. Both return highly significant results for the available statistical tests of the presence of phylogenetic signal. Yet both produce maximum parsimony trees which conflict with morphological analyses. This is also the case for distance-based trees and for the most realistic computable models for maximum likelihood analysis. The principal methods for dealing with noisy data are to weight sequence positions or types of changes (e.g. ignoring third codon positions in coding regions or giving transitions lower cost), omit some taxa or use constraint groups. In general, none of these approaches improves global concordance for the gastropod or arthropod studies. It is possible, however, by giving particular attention to the statistical significance of the results to test previously formulated hypotheses or to verify particular nodes within a suggested topology. Winning-sites tests and the use of the synthetic four-taxon comparisons implemented in Kumar's PHYLTEST program have been very useful in this context.
Gene rearrangements and translocations as phylogenetic markers for ancient evolutionary relationships. Are they really so rare?
Mark Dowton
Dept. Biology, Wollongong University, Wollongong, 2522; and Dept. Crop Protection, Waite Campus, Adelaide University, Glen Osmond, SA 5064
Ancient evolutionary relationships, such as those between the arthropod lineages, are perhaps the most difficult to decipher due to the rarity of retention of synapomorphic characters. Recent observations, gleaned from complete sequencing of the mitochondrial genomes of a few representative arthropods, suggest that the arrangement of mitochondrial genes may represent one such character; their relative positions appear highly conserved in the arthropods sampled thus far. However, very few mitochondrial genomes from arthropod taxa have been completely sequenced, making the claim for rarity tenuous. We have determined the arrangement of tRNA genes at a single mitochondrial locus among various members of the Hymenoptera. We find evidence of eight rearrangements among a group that has diverged over only 170 Myr (compared to the ca. 400 Myr of arthropod evolution). Further, we have identified two translocations, both involving the tRNA gene for histidine. Our data suggest that the rearrangement and translocatory mechanism may bias the historical pattern of changes. The hymenopteran mitochondrial genome promises to be a useful model for characterizing the mechanism for rearrangements and translocations.
New strategies for obtaining phylogenetic trees by maximum likelihood analysis
L.S. Jermiin
The Australian National University, Australia
The maximum-likelihood (ML) approach is a powerful tool for reconstructing molecular phylogenies. In conjunction with the Kishino-Hasegawa test, it allows direct comparison of alternative evolutionary hypotheses. A commonly occurring outcome is that several trees are not significantly different from the ML tree and there is thus residual uncertainty about the correct topology. In recognition of the fact that the trees that are not significantly different from the ML tree may contain some information about the correct tree topology, we present two new strategies for (1) identifying the ML tree and as many as possible of the other, less likely, trees and for (2) producing a consensus tree that is based on the ML tree and the other trees. The new maximum likelihood method is based on the stepwise addition approach but differs from that by the use of a much broader search window. The new consensus method is based on differential weighting of alternative tree topologies, with the tree-specific weights being based on statistics obtained from the ML analysis and the Kishino-Hasegawa test. Two computer programs, TrExML and TreeCons, have been developed to implement these new phylogenetic strategies.
Characterization of the control region and control region-like inserted copy in pythons
L.H. Rawlings1,2 and S.C. Donnellan1
1SA Museum,
2 Genetics Department, University of Adelaide
The gene arrangement of the mitochondrial genome has been shown to vary between vertebrate groups. In the mitochondrial genome of mammals, fish and amphibians, the control region is positioned between the cytochromeb/tRNATHY/tRNAPRO genes and the phenylalanine/12SrRNA genes. It has been shown that in some snakes there has been a duplication of the control region inserted into the WANCY region of the mitochondrial genome with an associated transfer of the proline and phenylalanine genes and a truncation or deletion of the tRNAPRO in its usual position (Kumazawa et al., 1996). In this study, we separately PCR amplified the control region and the control region-like sequences in pythons and compared the sequences. We found that there is an additional insertion in the 5' end of the control region of pythons that is not present in the viper, rattle snake and colubrid sequence of Kumazawa et al., nor in the Pacific boa, Candoia aspera. There is also a further insertion in Antaresia species. We found that whilst there is 4-5% variation between populations of carpet pythons, the control region-like duplication has maintained the fidelity of the original sequence.