As one of the major preys of many important economic fish species such as swordfish and dolphinfish in waters off estern Taiwan, flyingfish belongs to low-end consumers in the food chain with the function of maintaining the stability of the Kuroshio marine ecosystem. Hirundichthys oxycephalus is the primary component of flyingfish-egg fishery captures in the northeastern waters of Taiwan, and is also one of the dominant species of flyingfish in eastern waters of Taiwan. However, the significant drop of the flyingfish and flyingfish-egg catch from 2006 to 2007 and the effects on ecosystem and fishery caused major concern from the fishery sector and academic field. In order to manage this marine resource effectively, the phylogenetic relationships and population structure needed to be characterized first. In this study, the phylogenetic relationships of Hirundichthys oxycephalus of Northwestern Pacific was characterized based on the mitochondrial COI fragment. Totally 55 samples were collected between July, 2008 and November, 2010 in waters of Keelung, Ilan, Hualian, and Green Island. In addition, 12 more samples were obtained in Sebtember, 2009 from Tanegashima Island, and Yakushima Island of Japan. The DNA sequencing results of samples from Taiwan showed a total number of 29 haplotypes. The length of partial COI sequence was found to be 657 bp while the mean genetic distance was found to be 0.6%. In phylogenetic analyses, two major groups were identified in the phylogenetic trees by neighbor-joining and maximum-likelihood methods. The majority of "Keelung inshore group" came from Keelung and Ilan waters. The main population of "Kuroshio group" came from Green Island. The variation between two groups was found to be 61.75% by amova. The DNA sequencing results of samples from Japan showed a total number of 8 haplotypes. The length of partial COI sequences was found to be 657 bp with a mean genetic distance of 0.53%. In the phylogenetic tree, the samples from Japan were found to belong to "Kuroshio group". The variation between the two major groups was found to be 60% by amova. It was inferred that the differentiation of flyingfish into the two major groups in Taiwan was due to the flow pattern difference of Kuroshio in northeast waters of Taiwan. It was also inferred that phylogenetic similarity of the samples from Japan and the Kuroshio group was due to the distribution of both groups locating on the same path of the main current of Kuroshio. However, applying different distribution assumption may result in different conclusion such as one single stock hypothesis. Further studies will be needed to confirm the stock structure of the species.

Phylogenetic analysis can produce easily tens of thousands of equally plausible evolutionary trees. Consensus trees and topological distance matrices are often used to summarize the evolutionary relationships among the trees of interest. However, current approaches are not designed to analyze very large tree collections. In this dissertation, we present two fast algorithms— HashCS and HashRF —for analyzing large collections of evolutionary trees based on a novel hash table data structure, which provides a convenient and fast approach to store and access the bipartition information collected from the tree collections. Our HashCS algorithm is a fast ( ) technique for constructing consensus trees, where is the number of taxa and is the number of trees. By reprocessing the bipartition information in our hash table, HashCS constructs strict and majority consensus trees. In addition to a consensus algorithm, we design a fast topological distance algorithm called HashRF to compute the × Robinson-Foulds distance matrix, which requires ( ^ 2) running time. A RF distance matrix provides plenty of data-mining opportunities to help researchers understand the evolutionary relationships contained in their collection of trees. We also introduce a series of extensions based on HashRF to provide researchers with more convenient set of tools for analyzing their trees. We provide extensive experimentation regarding the practical performance of our hash-based algorithms across a diverse collection of biological and artificial trees. Our results show that both algorithms easily outperform existing consensus and RF matrix implementations. For example, on our biological trees, HashCS and HashRF are 1.8 and 100 times faster than PAUP*, respectively. We show two real-world applications of our fast hashing algorithms: (i) comparing phylogenetic heuristic implementations, and (ii) clustering and visualizing trees. In our first application, we design novel methods to compare the PaupRat and Rec-I-DCM3, two popular phylogenetic heuristics that use the Maximum Parsimony criterion, and show that RF distances are more effective than parsimony scores at identifying heterogeneity within a collection of trees. In our second application, we empirically show how to determine the distinct clusters of trees within large tree collections. We use two different techniques to identify distinct tree groups. Both techniques show that partitioning the trees into distinct groups and summarizing each group separately is a better representation of the data. Additional benefits of our approach are better consensus trees as well as insightful information regarding the convergence behavior of phylogenetic heuristics. Our fast hash-based algorithms provide scientists with a very powerful tools for analyzing the relationships within their large phylogenetic tree collections in new and exciting ways. Our work has many opportunities for future work including detecting convergence and designing better heuristics. Furthermore, our hash tables have lots of potential future extensions. For example, we can also use our novel hashing structure to design algorithms for computing other distance metrics such as Nearest Neighbor Interchange (NNI), Subtree Pruning and Regrafting (SPR), and Tree Bisection and Reconnection (TBR) distances.

Purple nonsulfur bacteria are a group of extraordinary metabolic diverse bacteria. They can grow photoautotrophically, photoheterotrophically , chemoheterotrophically or chemoautotrophically. Under various conditions, they can enjoy exceptional flexibility within each of these modes of metabolism. Due to the special physical characteristics properties, they had attracted scientist¡¦s attention in resent years. These bacteria are widely distributed in nature such as lakes, water ponds, coastal lagoons or high concentration organic waste lagoons. Lotus Pond, located in northern Kaohsiung City, is a serious eutrophied artificial lake. Because of receiving sufficient light and having been polluted by significant amounts of soluble organic matter, the ecology of the lake is suitable for the growth of purple nonsulfur bacteria. In the study, the lake water and sediments by using a Winograsdsky column, we successfully isolated 16 strains bacteria from the Lotus Pond. We also amplified the 16S-rDNA fragments of these strains by PCR and sequenced these PCR products, then aligned these sequences with the data of GeneBank. We affirmed that the 16 isolated strains belong to purple nonsulfur bacteria. From phylogenetic analysis, these 16 strains belong to the following three groups of bacteria: Rhodopseudomonas palustris, Rubrivivax gelatinosus, and Rhodobacter sphaeroides. Characteristic studies of these strains, we found that all isolated strains are Gram negative bacteria and contain bacteriochlorophyll a. The strains that belong to R. palustris and R. sphaeroides group can use several different types of short chain organic acid as their carbon source and have denitrification ability. However, only the strains belong to R. palustris group are able to use the aromatic compound benzoate. From salt tolerant studies, we found the strains in R. sphaeroides group can grow well in 3% NaCl, and both R. palustris and R. gelatinosus group can only grow in 1% NaCl.

Tzukuan is an HBV-, HCV-, HDV- endemic township in southern Taiwan. Based on a mass screening on 2909 residents age of 45 years or more in 1997, the prevalence rates of HBsAg and anti-HCV were 12.8% and 41.6% respectively. Of HBsAg carriers, 15.3% were positive for anti-HDV. Tzukaun was divided into coastal area and inland area. The prevalence of anti-HCV of coastal area was two times higher than that of inland area (61.4% v.s. 29.1%) and genotype 1b and 2a are the main two subtypes. We wish to find the causes of discrepancy in these nearby areas by phylogenetic analysis. Stratified by the living areas, coastal or inland, 27 samples were picked up (ingroup). HCV sequence of NS5B region could be detected by RT-PCR then a nested PCR in eight males and ninteen females with mean age of 54.8 years old (range: 45-70). None of these 27 residents came from the same family. Another 10 HCV infected persons whose living townships also in southern Taiwan but other than Tzukuan were enrolled as local controls. From GenBank, 30 different HCV isolates were included. Phylogenic analysis unequivocally confirmed the simultaneous spread of two different HCV strains in this township clusters according to their subtypes were noted. A trend of the spreading from coastal to land area or an ultra-aggregation phynomenon which according to their living area, as we suspected, were not noted between Tzukuan¡¦s residents. In ingroup, the short genetic distance between the isolates of C hepatitis virus which came from different villages might be caused from the wide-spreading of HCV in this endemic area (the maximal and minimal genetic distance in 1b or 2a isolates are 0.0869 vs. 0.0098 and 0.0996 vs. 0.0334). Besides, according to the contacting history to foreigner by our aborigine tribes, from genebank, all isolates from different countries were included and three possible origins of HCV genotype 1b were noted in Tzukuan. All these findings might be caused from frequently HCV inflow in this endemic area and wide-spreading of HCV between different countries.

Recently, the stochastic approximation Monte Carlo algorithm has been proposed by Liang et al. (2005) as a general-purpose stochastic optimization and simulation algorithm. An annealing version of this algorithm was developed for real small protein folding problems. The numerical results indicate that it outperforms simulated annealing and conventional Monte Carlo algorithms as a stochastic optimization algorithm. We also propose one method for the use of secondary structures in protein folding. The predicted protein structures are rather close to the true structures. Phylogenetic trees have been used in biology for a long time to graphically represent evolutionary relationships among species and genes. An understanding of evolutionary relationships is critical to appropriate interpretation of bioinformatics results. The use of the sequential structure of phylogenetic trees in conjunction with stochastic approximation Monte Carlo was developed for phylogenetic tree reconstruction. The numerical results indicate that it has a capability of escaping from local traps and achieving a much faster convergence to the global likelihood maxima than other phylogenetic tree reconstruction methods, such as BAMBE and MrBayes.

Reticuloendotheliosis virus (REV) is a retrovirus that causes a neoplastic disease in a wide range of avian hosts including chickens, turkeys, and ducks. In 1993, REV was detected in the endangered Attwater's prairie chicken (Tympanachus cupido attwateri), a subspecies of Tympanachus cupido. Subsequent infections of this prairie chicken have been identified at captive breeding facilities throughout Texas. The implications of these infections have severely hindered repopulation efforts by these facilities. This study focused on investigating REV infection of captive Attwater'€™s prairie chicken in order to better understand the disease affecting these endangered birds. The overall objective was to develop a means of eliminating this threat to the repopulation of the Attwater's prairie chicken. Several aspects of virus infection were investigated. Reagents capable of recognizing prairie chicken IgY and viral gag polypeptides were developed for use in assays for detection of antibody responses and titration of viral concentrations. Sequencing data of genomes collected from isolates of Texas prairie chickens and domestic chickens, as well as three REV prototype viruses, were compared to determine relationships among strains and identify the potential origin of the REV infecting Attwater'€™s prairie chicken. Additionally, a flow cytometry technique of segregating the lymphocyte population from peripheral blood mononuclear cells (PBMC) using a pan leukocyte monoclonal antibody was developed to more accurately measure changes within lymphocyte populations. This technique combined with intracellular labeling was used to deduce the target cells of REV infection. A nested polymerase chain reaction (PCR) test was developed for greater sensitivity in detecting infection in birds than the previous method of single amplification PCR. This greater sensitivity results in earlier identification of the virus in infected birds, which allows for earlier removal of infected birds to minimize transmission of the virus throughout the flock. The sensitivity of the nested PCR diagnostic test was determined in a dose response pathogenesis study, which was conducted on hybrid greater/Attwater's prairie chicken to observe the experimental development of disease in these birds. Finally, a vaccine was developed using plasmid DNA with REV encoded genes and tested on naturally infected prairie chickens to determine its efficacy in reducing viral load. Although no reduction in viral load was detected, the vaccine may be effective in providing prophylactic protection in future studies.

Ever since Darwin claimed that expanding savannas were the driving force behind humanity's divergence from other apes, our understanding of human evolution has been inextricably linked to the environmental context in which our ancestors evolved. This dissertation explores various aspects of the use of one method of paleoenvironmental reconstruction -- bovid ecomorphology -- and provides new data on paleoenvironmental conditions in the Omo Shungura Formation (Plio-Pleistocene, Ethiopia). Chapter 2 uses phylogenetic simulations to explore the performance of Discriminant Function Analysis (DFA) on simulated ecomorphological data containing phylogenetic signal. DFA is shown to "over-perform" in situations in which predicted and predictor variables both contain phylogenetic signal. Phylogenetic Generalized Least Squares (PGLS) is shown to be a very useful technique for explicitly testing functional hypotheses in ecomorphology while controlling for phylogenetic signal and body size. Chapter 3 presents a functional analysis of the bovid astragalus, which is one of the most commonly preserved bones in the fossil record. Several functional hypotheses linking habitat-specific locomotor performance with the morphology of the astragalus are tested using PGLS. Strong support is found for three of these hypotheses. Thus, the astragalus is shown to be a useful ecomorphological predictor element, a point that is confirmed by the DFA analyses in Chapter 4. Chapter 5 provides new paleoenvironmental data on the Omo Shungura Formation based on habitat reconstructions from astragalar ecomorphology in addition to dietary reconstructions based on dental mesowear. Astragalar data point to a major environmental shift beginning ~2.58 Ma, which is later in time compared with some prior habitat reconstructions using different methods. Furthermore, astragalar data show environmental fluctuations of similar magnitude later in the sequence. Mesowear data on the Shungura Tragelaphini do not offer evidence for any significant grazing adaptation, in spite of relatively high carbon isotope signatures reported based on studies of tooth enamel. These data raise questions regarding the diet of fossil Tragalephini. / text

Climate change has been invoked to explain patterns of speciation, extinction and biogeographic change over time, however it can be a difficult hypothesis to test empirically. One area of particular interest is climate change in the African Neogene, linked with the origin of hominins. A perfect model clade to test these hypotheses is the papionin monkeys, a diverse group (both extinct and extant), represented by an excellent fossil record. I describe new fossil papionin specimens from Coopers Cave, South Africa, and redescribe and discuss some previously known fossil material. This rich data set provides a necessary deep-time perspective, and, in conjunction with independently generated data, can be used to test hypotheses related to climatic and geological events (such as increasing late Pleistocene aridity and persistence of forest refugia) that may be directly linked to patterns of speciation and biogeographic distribution in the fossil record and in living species. Testing these hypotheses requires a robust phylogenetic hypothesis. I collected morphological character data for a species-level phylogenetic analysis of the papionin clade in order to reconstruct the phylogeny of the group. My analysis found that the living species Theropithecus gelada is nested within extinct theropiths, and is primitive relative to the Pleistocene taxa Theropithecus darti, T. oswaldi and T. leakeyi. Also falling within the theropith lineage are the early Pliocene taxon Pliopapio, the South African taxa Dinopithecus and Gorgopithecus, and two species whose relationships were uncertain until my analysis. “Papio” quadratirostris and “Papio” baringensis are nested within the theropiths, and should be referred to the genus Theropithecus. Biogeographic analysis demonstrates that papionin monkeys share a similar pattern with other Neogene African mammals; they first disperse out of Africa during the mid-Miocene, return to Africa by the late Miocene and undergo a series of vicariant speciation events and range restriction to central Africa, but disperse out into eastern and southern Africa by the Pleistocene. These speciation and dispersal events are tightly correlated with global climatic and tectonic changes.

In vitro gut models provide several advantages over in vivo models for the study of the human gut microbiota. However, because communities developed in these models are simplified simulations of the in vivo environment it is necessary to characterize the reproducibility, repeatability and stability of cultured communities. We also need to broadly define the differences between in vitro consortia and the communities from which they are derived. In this study we characterized and validated a twin-vessel (independent, identical) single-stage chemostat model of the human distal gut. Samples were analyzed using a molecular fingerprinting technique (Denaturing Gradient Gel Electrophoresis) to compare and monitor changes in the overall structure of the communities while a phylogenetic microarray (Human Intestinal Tract Chip) was used to obtain phylogenetic information. We found that twin-vessels inoculated with feces developed and maintained diverse communities that reached stable compositions by at most 36 days post-inoculation. Communities were enriched in Bacteroidetes but not Clostridium cluster XIVa, Bacilli or other Firmicutes relative to the fecal inocula. Vessels were very reproducible when inoculated with identical fecal inocula, less similar when inoculated with consecutive fecal donations from the same donor, and maintained donor-specific identities when inoculated with feces from different donors. Norepinephrine exposure (undefined perturbation) did not appear to have a substantial effect on the structure of chemostat communities, while clindamycin treatment (defined perturbation) caused large changes in the structure of chemostat communities. Packed-column biofilm reactors incorporated a simulated mucosal environment into our chemostat system, allowing us to simultaneously culture biologically relevant planktonic and biofilm communities that were complex, reproducible, and distinct. Defined communities were comparable to fecal communities at the phylum/class-level but established stable compositions more rapidly. While it was difficult to assess the persistence of synthetic stool in a healthy fecal chemostat community (+/- antibiotic perturbation), mixing communities from two donors resulted in a mixed community that more closely resembled one donor over the other. Although future experimentation is required, the results presented here show our twin-vessel single-stage chemostat model represents a valid simulation of the human distal gut environment and can support complex, representative microbial communities ideal for experimental manipulation. / Canadian Institutes of Health Research; Ontario Ministry of Agriculture, Food and Rural Affairs; Ontario Ministry of Research and Innovation; Canada Foundation for Innovation; Ontario Ministry of Training, Colleges and Universities

Brachyglottis (Asteraceae) is a genus of approximately 30 species in the Brachyglottidinae, a recently recognised sub-tribe of tribe Senecioneae. Within Brachyglottis is a clade of five species of rosette-forming herbs: B. bellidioides, B. haastii, B. lagopus, B. southlandica and B. traversii. A sixth species, B. saxifragoides, has recently been synonymised with B. lagopus. The rosette-Brachyglottis have historically been recognised as a taxonomically problematic group because species overlap in both morphology and geographical distribution. A recent molecular study of rosette-Brachyglottis using AFLP data indicated that genetic distances among populations of rosette-Brachyglottis in the South Island appear to be correlated with geographical distance between populations rather than taxonomic identification. This is problematic as the currently described rosette-Brachyglottis species have overlapping ranges which implicitly hypothesises reproductive barriers other than geographic distance. We conducted an investigation into the species delimitation of rosette-Brachyglottis with the aim of answering two related questions: Does the current delimitation of rosette-Brachyglottis accurately reflect patterns of genetic similarity? Do the patterns of genetic structure in rosette-Brachyglottis support the presence of multiple biological species? A total of 46 populations of rosette-Brachyglottis were represented in this study. Herbarium specimens collected from these populations were identified following the taxonomic treatment of Allan (1961). Twenty one discrete and numerical morphological characters were measured from herbarium specimens including those collected for this study and previously collected herbarium specimens. Morphological dissimilarity of 354 herbarium specimens was investigated by performing a PCoA on Gower’s pairwise morphological distances among individuals. The pattern of genetic similarity was explored using DNA fragment length variation in nine markers for 273 individuals and this resulted in a total 177 unique alleles. Bayesian clustering analysis was performed on this data set using STRUCTURE, in addition, pairwise genetic distances were calculated among individuals and populations using Jaccard and Nei’s dissimilarity coefficient’s respectively. Jaccard genetic distances among individuals were analysed using PCoA and Nei’s genetic distances among populations were analysed using a Neighbour-Net analysis. The relationship between pairwise genetic and geographic distances among populations was analysed using a combination of linear regression and a Mantel Test. The pattern of morphological similarities among specimens was generally congruent with the currently delimited species in rosette-Brachyglottis. However, many morphologically intermediate specimens confound the recognition of distinct morphological entities. Comparison of patterns of genetic similarity and the current morphologically-based species delimitation showed that the delimitation does not accurately reflect the genetic structure of rosette-Brachyglottis. Furthermore, patterns of genetic dissimilarity did not indicate discrete genetic groups at the individual or population levels. The finding of incongruence between patterns of genetic and morphological similarity and absence of morphologically or genetically discrete groups suggests that rosette-Brachyglottis are best considered a single, yet morphologically diverse, biological species. In addition genetic structure within this species appears to be primarily driven by geographical isolation.