@phdthesis{Fasemore2023,
  author    = {Fasemore, Akinyemi Mandela},
  title     = {Genomic and internet based analysis of \(Coxiella\) \(burnetii\)},
  doi       = {10.25972/OPUS-29663},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-296639},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2023},
  abstract  = {Coxiella burnetii, a Gram negative obligate intracellular bacterium, is the causative agent of Q fever. It has a world wide distribution and has been documented to be capable of causing infections in several domestic animals, livestock species, and human beings. Outbreaks of Q fever are still being observed in livestock across animal farms in Europe, and primary transmission to humans still oc- curs especially in animal handlers. Public health authorities in some countries like Germany are required by law to report human acute cases denoting the significance of the challenge posed by C. burnetii to public health. In this thesis, I have developed a platform alongside methods to address the challenges of genomic analyses of C. burnetii for typing purposes. Identification of C. burnetii isolates is an important task in the laboratory as well as in the clinics and genotyping is a reliable method to identify and characterize known and novel isolates. Therefore, I designed and implemented several methods to facilitate the genotyping analyses of C. burnetii genomes in silico via a web platform. As genotyping is a data intensive process, I also included additional features such as visualization methods and databases for interpretation and storage of obtained results. I also developed a method to profile the resistome of C. burnetii isolates using a machine learning approach. Data about antibiotic resistance in C. burnetii are scarce majorly due to its lifestyle and the difficulty of cultivation in laboratory media. Alternative methods that rely on homology identification of resistance genes are also inefficient in C. burnetii, hence, I opted for a novel approach that has been shown to be promising in other bacteria species. The applied method relied on an artificial neural network as well as amino acid composition of position specific scoring matrix profile for feature extraction. The resulting model achieved an accuracy of ≈ 0.96 on test data and the overall performance was significantly higher in comparison to existing models. Finally, I analyzed two new C. burnetii isolates obtained from an outbreak in Germany, I compared the genome to the RSA 493 reference isolate and found extensive deletions across the genome landscape. This work has provided a new digital infrastructure to analyze and character- ize C. burnetii genomes that was not in existence before and it has also made a significant contribution to the existing information about antibiotic resistance genes in C. burnetii.},
  language  = {en}
}
@phdthesis{Bischler2018,
  author    = {Bischler, Thorsten David},
  title     = {Data mining and software development for RNA-seq-based approaches in bacteria},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-166108},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2018},
  abstract  = {RNA sequencing (RNA-seq) has in recent years become the preferred method for gene expression analysis and whole transcriptome annotation. While initial RNA-seq experiments focused on eukaryotic messenger RNAs (mRNAs), which can be purified from the cellular ribonucleic acid (RNA) pool with relative ease, more advanced protocols had to be developed for sequencing of microbial transcriptomes. The resulting RNA-seq data revealed an unexpected complexity of bacterial transcriptomes and the requirement for specific analysis methods, which in many cases is not covered by tools developed for processing of eukaryotic data. The aim of this thesis was the development and application of specific data analysis methods for different RNA-seq-based approaches used to gain insights into transcription and gene regulatory processes in prokaryotes. The differential RNA sequencing (dRNA-seq) approach allows for transcriptional start site (TSS) annotation by differentiating between primary transcripts with a 5'-triphosphate (5'-PPP) and processed transcripts with a 5'-monophosphate (5'-P). This method was applied in combination with an automated TSS annotation tool to generate global trancriptome maps for Escherichia coli (E. coli) and Helicobacter pylori (H. pylori). In the E. coli study we conducted different downstream analyses to gain a deeper understanding of the nature and properties of transcripts in our TSS map. Here, we focused especially on putative antisense RNAs (asRNAs), an RNA class transcribed from the opposite strand of known protein-coding genes with the potential to regulate corresponding sense transcripts. Besides providing a set of putative asRNAs and experimental validation of candidates via Northern analysis, we analyzed and discussed different sources of variation in RNA-seq data. The aim of the H. pylori study was to provide a detailed description of the dRNA-seq approach and its application to a bacterial model organism. It includes information on experimental protocols and requirements for data analysis to generate a genome-wide TSS map. We show how the included TSS can be used to identify and analyze transcriptome and regulatory features and discuss challenges in terms oflibrary preparation protocols, sequencing platforms, and data analysis including manual and automated TSS annotation. The TSS maps and associated transcriptome data from both H. pylori and E. coli were made available for visualization in an easily accessible online browser. Furthermore, a modified version of dRNA-seq was used to identify transcriptome targets of the RNA pyrophosphohydrolase (RppH) in H. pylori. RppH initiates 5'-end-dependent degradation of transcripts by converting the 5'-PPP of primary transcripts to a 5'-P. I developed an analysis method, which uses data from complementary DNA (cDNA) libraries specific for transcripts carrying a 5'-PPP, 5'-P or both, to specifically identify transcripts modified by RppH. For this, the method assessed the 5'-phosphorylation state and cellular concentration of transcripts in rppH deletion in comparison to strains with the intact gene. Several of the identified potential RppH targets were further validated via half-life measurements and quantification of their 5'-phosphorylation state in wild-type and mutant cells. Our findings suggest an important role for RppH in post-transcriptional gene regulationin H. pylori and related organisms. In addition, we applied two RNA-seq -based approaches, RNA immunoprecipitation followed by sequencing (RIP-seq) and cross-linking immunoprecipitation followed by sequencing (CLIP-seq), to identify transcripts bound by Hfq and CsrA, two RNA-binding proteins (RBPs) with an important role in post-transcriptional regulation. For RIP-seq -based identification of CsrA binding regions in Campylobacter jejuni(C. jejuni), we used annotation-based analysis and, in addition, a self-developed peak calling method based on a sliding window approach. Both methods revealed flaA mRNA, encoding the major flagellin, as the main target and functional analysis of identified targets showed a significant enrichment of genes involved in flagella biosynthesis. Further experimental analysis revealed the role of flaA mRNA in post-transcriptional regulation. In comparison to RIP-seq, CLIP-seq allows mapping of RBP binding sites with a higher resolution. To identify these sites an approach called "block-based peak calling" was developed and resulting peaks were used to identify sequence and structural constraints required for interaction of Hfq and CsrA with Salmonella transcripts. Overall, the different RNA-seq-based approaches described in this thesis together with their associated analyis pipelines extended our knowledge on the transcriptional repertoire and modes of post-transcriptional regulation in bacteria. The global TSS maps, including further characterized asRNA candidates, putative RppH targets, and identified RBP interactomes will likely trigger similar global studies in the same or different organisms or will be used as a resource for closer examination of these features.},
  subject      = {Bakterien},
  language  = {en}
}
@phdthesis{Sharan2017,
  author    = {Sharan, Malvika},
  title     = {Bio-computational identification and characterization of RNA-binding proteins in bacteria},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-153573},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2017},
  abstract  = {RNA-binding proteins (RBPs) have been extensively studied in eukaryotes, where they post-transcriptionally regulate many cellular events including RNA transport, translation, and stability. Experimental techniques, such as cross-linking and co-purification followed by either mass spectrometry or RNA sequencing has enabled the identification and characterization of RBPs, their conserved RNA-binding domains (RBDs), and the regulatory roles of these proteins on a genome-wide scale. These developments in quantitative, high-resolution, and high-throughput screening techniques have greatly expanded our understanding of RBPs in human and yeast cells. In contrast, our knowledge of number and potential diversity of RBPs in bacteria is comparatively poor, in part due to the technical challenges associated with existing global screening approaches developed in eukaryotes. Genome- and proteome-wide screening approaches performed in silico may circumvent these technical issues to obtain a broad picture of the RNA interactome of bacteria and identify strong RBP candidates for more detailed experimental study. Here, I report APRICOT ("Analyzing Protein RNA Interaction by Combined Output Technique"), a computational pipeline for the sequence-based identification and characterization of candidate RNA-binding proteins encoded in the genomes of all domains of life using RBDs known from experimental studies. The pipeline identifies functional motifs in protein sequences of an input proteome using position-specific scoring matrices and hidden Markov models of all conserved domains available in the databases and then statistically score them based on a series of sequence-based features. Subsequently, APRICOT identifies putative RBPs and characterizes them according to functionally relevant structural properties. APRICOT performed better than other existing tools for the sequence-based prediction on the known RBP data sets. The applications and adaptability of the software was demonstrated on several large bacterial RBP data sets including the complete proteome of Salmonella Typhimurium strain SL1344. APRICOT reported 1068 Salmonella proteins as RBP candidates, which were subsequently categorized using the RBDs that have been reported in both eukaryotic and bacterial proteins. A set of 131 strong RBP candidates was selected for experimental confirmation and characterization of RNA-binding activity using RNA co-immunoprecipitation followed by high-throughput sequencing (RIP-Seq) experiments. Based on the relative abundance of transcripts across the RIP-Seq libraries, a catalogue of enriched genes was established for each candidate, which shows the RNA-binding potential of 90\% of these proteins. Furthermore, the direct targets of few of these putative RBPs were validated by means of cross-linking and co-immunoprecipitation (CLIP) experiments. This thesis presents the computational pipeline APRICOT for the global screening of protein primary sequences for potential RBPs in bacteria using RBD information from all kingdoms of life. Furthermore, it provides the first bio-computational resource of putative RBPs in Salmonella, which could now be further studied for their biological and regulatory roles. The command line tool and its documentation are available at https://malvikasharan.github.io/APRICOT/.},
  language  = {en}
}
@phdthesis{Horn2017,
  author    = {Horn, Hannes},
  title     = {Analysis and interpretation of (meta-)genomic data from host-associated microorganisms},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-152035},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2017},
  abstract  = {Host-microbe interactions are the key to understand why and how microbes inhabit specific environments. With the scientific fields of microbial genomics and metagenomics, evolving on an unprecedented scale, one is able to gain insights in these interactions on a molecular and ecological level. The goal of this PhD thesis was to make (meta-)genomic data accessible, integrate it in a comparative manner and to gain comprehensive taxonomic and functional insights into bacterial strains and communities derived from two different environments: the phyllosphere of Arabidopsis thaliana and the mesohyl interior of marine sponges. This thesis focused first on the de novo assembly of bacterial genomes. A 5-step protocol was developed, each step including a quality control. The examination of different assembly software in a comparative way identified SPAdes as most suitable. The protocol enables the user to chose the best tailored assembly. Contamination issues were solved by an initial filtering of the data and methods normally used for the binning of metagenomic datasets. This step is missed in many published assembly pipelines. The described protocol offers assemblies of high quality ready for downstream analysis. Subsequently, assemblies generated with the developed protocol were annotated and explored in terms of their function. In a first study, the genome of a phyllosphere bacterium, Williamsia sp. ARP1, was analyzed, offering many adaptions to the leaf habitat: it can deal with temperature shifts, react to oxygen species, produces mycosporins as protection against UV-light, and is able to uptake photosynthates. Further, its taxonomic position within the Actinomycetales was infered from 16S rRNA and comparative genomics showing the close relation between the genera Williamsia and Gordonia. In a second study, six sponge-derived actinomycete genomes were investigated for secondary metabolism. By use of state-of-the-art software, these strains exhibited numerous gene clusters, mostly linked to polykethide synthases, non-ribosomal peptide synthesis, terpenes, fatty acids and saccharides. Subsequent predictions on these clusters offered a great variety of possible produced compounds with antibiotic, antifungal or anti-cancer activity. These analysis highlight the potential for the synthesis of natural products and the use of genomic data as screening toolkit. In a last study, three sponge-derived and one seawater metagenomes were functionally compared. Different signatures regarding the microbial composition and GC-distribution were observed between the two environments. With a focus on bacerial defense systems, the data indicates a pronounced repertoire of sponge associated bacteria for bacterial defense systems, in particular, Clustered Regularly Interspaced Short Palindromic Repeats, restriction modification system, DNA phosphorothioation and phage growth limitation. In addition, characterizing genes for secondary metabolite cluster differed between sponge and seawater microbiomes. Moreover, a variety of Type I polyketide synthases were only found within the sponge microbiomes. With that, metagenomics are shown to be a useful tool for the screening of secondary metabolite genes. Furthermore, enriched defense systems are highlighted as feature of sponge-associated microbes and marks them as a selective trait.},
  subject      = {Bakterien},
  language  = {en}
}
@phdthesis{ZeeshangebMajeed2014,
  author    = {Zeeshan [geb. Majeed], Saman},
  title     = {Implementation of Bioinformatics Methods for miRNA and Metabolic Modelling},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-102900},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2014},
  abstract  = {Dynamic interactions and their changes are at the forefront of current research in bioinformatics and systems biology. This thesis focusses on two particular dynamic aspects of cellular adaptation: miRNA and metabolites. miRNAs have an established role in hematopoiesis and megakaryocytopoiesis, and platelet miRNAs have potential as tools for understanding basic mechanisms of platelet function. The thesis highlights the possible role of miRNAs in regulating protein translation in platelet lifespan with relevance to platelet apoptosis and identifying involved pathways and potential key regulatory molecules. Furthermore, corresponding miRNA/target mRNAs in murine platelets are identified. Moreover, key miRNAs involved in aortic aneurysm are predicted by similar techniques. The clinical relevance of miRNAs as biomarkers, targets, resulting later translational therapeutics, and tissue specific restrictors of genes expression in cardiovascular diseases is also discussed. In a second part of thesis we highlight the importance of scientific software solution development in metabolic modelling and how it can be helpful in bioinformatics tool development along with software feature analysis such as performed on metabolic flux analysis applications. We proposed the "Butterfly" approach to implement efficiently scientific software programming. Using this approach, software applications were developed for quantitative Metabolic Flux Analysis and efficient Mass Isotopomer Distribution Analysis (MIDA) in metabolic modelling as well as for data management. "LS-MIDA" allows easy and efficient MIDA analysis and, with a more powerful algorithm and database, the software "Isotopo" allows efficient analysis of metabolic flows, for instance in pathogenic bacteria (Salmonella, Listeria). All three approaches have been published (see Appendices).},
  subject      = {miRNS},
  language  = {en}
}
@phdthesis{Blenk2007,
  author    = {Blenk, Steffen},
  title     = {Bioinformatical analysis of B-cell lymphomas},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-27421},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2007},
  abstract  = {Background: The frequency of the most observed cancer, Non Hodgkin Lymphoma (NHL), is further rising. Diffuse large B-cell lymphoma (DLBCL) is the most common of the NHLs. There are two subgroups of DLBCL with different gene expression patterns: ABC ("Activated B-like DLBCL") and GCB ("Germinal Center B-like DLBCL"). Without therapy the patients often die within a few months, the ABC type exhibits the more aggressive behaviour. A further B-cell lymphoma is the Mantle cell lymphoma (MCL). It is rare and shows very poor prognosis. There is no cure yet. Methods: In this project these B-cell lymphomas were examined with methods from bioinformatics, to find new characteristics or undiscovered events on the molecular level. This would improve understanding and therapy of lymphomas. For this purpose we used survival, gene expression and comparative genomic hybridization (CGH) data. In some clinical studies, you get large data sets, from which one can reveal yet unknown trends. Results (MCL): The published proliferation signature correlates directly with survival. Exploratory analyses of gene expression and CGH data of MCL samples (n=71) revealed a valid grouping according to the median of the proliferation signature values. The second axis of correspondence analysis distinguishes between good and bad prognosis. Statistical testing (moderate t-test, Wilcoxon rank-sum test) showed differences in the cell cycle and delivered a network of kinases, which are responsible for the difference between good and bad prognosis. A set of seven genes (CENPE, CDC20, HPRT1, CDC2, BIRC5, ASPM, IGF2BP3) predicted, similarly well, survival patterns as proliferation signature with 20 genes. Furthermore, some bands could be associated with prognosis in the explorative analysis (chromosome 9: 9p24, 9p23, 9p22, 9p21, 9q33 and 9q34). Results (DLBCL): New normalization of gene expression data of DLBCL patients revealed better separation of risk groups by the 2002 published signature based predictor. We could achieve, similarly well, a separation with six genes. Exploratory analysis of gene expression data could confirm the subgroups ABC and GCB. We recognized a clear difference in early and late cell cycle stages of cell cycle genes, which can separate ABC and GCB. Classical lymphoma and best separating genes form a network, which can classify and explain the ABC and GCB groups. Together with gene sets which identify ABC and GCB we get a network, which can classify and explain the ABC and GCB groups (ASB13, BCL2, BCL6, BCL7A, CCND2, COL3A1, CTGF, FN1, FOXP1, IGHM, IRF4, LMO2, LRMP, MAPK10, MME, MYBL1, NEIL1 and SH3BP5; Altogether these findings are useful for diagnosis, prognosis and therapy (cytostatic drugs).},
  subject      = {Bioinformatik},
  language  = {en}
}
@phdthesis{Pils2005,
  author    = {Pils, Birgit},
  title     = {Insights into the evolution of protein domains give rise to improvements of function prediction},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-16805},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2005},
  abstract  = {The growing number of uncharacterised sequences in public databases has turned the prediction of protein function into a challenging research field. Traditional annotation methods are often error-prone due to the small subset of proteins with experimentally verified function. Goal of this thesis was to analyse the function and evolution of protein domains in order to understand molecular processes in the cell. The focus was on signalling domains of little understood function, as well as on functional sites of protein domains in general. Glucosaminidases (GlcNAcases) represent key enzymes in signal transduction pathways. Together with glucosamine transferases, they serve as molecular switches, similar to kinases and phosphatases. Little was known about the molecular function and structure of the GlcNAcases. In this thesis, the GlcNAcases were identified as remote homologues of N-acetyltransferases. By comparing the homologous sequences, I was able to predict functional sites of the GlcNAcase family and to identify the GlcNAcases as the first family member of the acetyltransferase superfamily with a distinct catalytic mechanism, which is not involved in the transfer of acetyl groups. In a similar approach, the sensor domain of a plant hormone receptor was studied. I was able to predict putative ligand-binding sites by comparing evolutionary constraints in functionally diverged subfamilies. Most of the putative ligand-binding sites have been experimentally confirmed in the meantime. Due to the importance of enzymes involved in cellular signalling, it seems impossible to find substitutions of catalytic amino acids that turn them catalytically inactive. Nevertheless, by scanning catalytic positions of the protein tyrosine phosphatase families, I found many inactive domains among single domain and tandem domain phosphatases in metazoan proteomes. In addition, I found that inactive phosphatases are conserved throughout evolution, which led to the question about the function of these catalytically inactive phosphatase domains. An analysis of evolutionary site rates of amino acid substitutions revealed a cluster of conserved residues in the apparently redundant domain of tandem phosphatases. This putative regulatory center might be responsible for the experimentally verified dimerization of the active and inactive domain in order to control the catalytic activity of the active phosphatase domain. Moreover, I detected a subgroup of inactive phosphatases, which presumably functions in substrate recognition, based on different evolutionary site rates within the phosphatase family. The characterization of these new regulatory modules in the phosphatase family raised the question whether inactivation of enzymes is a more general evolutionary mechanism to enlarge signalling pathways and whether inactive domains are also found in other enzyme families. A large-scale analysis of substitutions at catalytic positions of enzymatic domains was performed in this work. I identified many domains with inactivating substitutions in various enzyme families. Signalling domains harbour a particular high occurrence of catalytically inactive domains indicating that these domains have evolved to modulate existing regulatory pathways. Furthermore, it was shown that inactivation of enzymes by single substitutions happened multiple times independently in evolution. The surprising variability of amino acids at catalytic positions was decisive for a subsequent analysis of the diversity of functional sites in general. Using functional residues extracted from structural complexes I could show that functional sites of protein domains do not only vary in their type of amino acid but also in their structural location within the domain. In the process of evolution, protein domains have arisen from duplication events and subsequently adapted to new binding partners and developed new functions, which is reflected in the high variability of functional sites. However, great differences exist between domain families. The analysis demonstrated that functional sites of nuclear domains are more conserved than functional sites of extracellular domains. Furthermore, the type of ligand influences the degree of conservation, for example ion binding sites are more conserved than peptide binding sites. The work presented in this thesis has led to the detection of functional sites in various protein domains involved in signalling pathways and it has resulted in insights into the molecular function of those domains. In addition, properties of functional sites of protein domains were revealed. This knowledge can be used in the future to improve the prediction of protein function and to identify functional sites of proteins.},
  subject      = {Dom{\"a}ne <Biochemie>},
  language  = {en}
}
@phdthesis{FadlElMola2003,
  author    = {Fadl El Mola, Faisal Mohamed},
  title     = {Bioinformatic and molecular approaches for the analysis of the retinal pigment epithelium (RPE) transcriptome},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-6877},
  school      = {Universit{\"a}t W{\"u}rzburg},
  year      = {2003},
  abstract  = {There is substantial interest in the identification of genes underlying susceptibility to complex human diseases because of the potential utility of such genes in disease prediction and therapy. The complex age-related macular degeneration (AMD) is a prevalent cause of legal blindness in industrialized countries and predominantly affects the elderly population over 75 years of age. Although vision loss in AMD results from photoreceptor cell death in the central retina, the initial pathogenesis likely involves processes in the retinal pigment epithelium (RPE) (Liang and Godley, 2003). The goal of the current study was to identify and characterize genes specifically or abundantly expressed in the RPE in order to determine more comprehensively the transcriptome of the RPE. In addition, our aim was to assess the role of these genes in AMD pathogenesis. Towards this end, a bovine cDNA library enriched for RPE transcripts was constructed in-house using a PCR-based suppression subtractive hybridization (SSH) technique (Diatchenko et al., 1996, 1999), which normalizes for sequence abundance and achieves high enrichment for differentially expressed genes. CAP3 (Huang and Madan, 1999) was used to assemble the high quality sequences of all the 2379 ESTs into clusters or singletons. 1.2\% of the 2379 RPE-ESTs contains vector sequences and was excluded from further analysis. 5\% of the RPE-ESTs showed homology to multipe chromosomes and were not included in further assembly process. The rest of the ESTs (2245) were assembled into 175 contigs and 509 singletons, which revealed approximately 684 unique genes in the dataset. Out of the 684, 343 bovine RPE transcripts did not align to their human orthologues. A large fraction of clones were shown to include a considerable 3´untranslated regions of the gene that are not conserved between bovine and human. It is the coding regions that can be conserved between bovine and human and not the 3' UTR (Sharma et al., 2002). Therefore, more sequencing from the cDNA library with reclustering of those 343 ESTs together with continuous blasting might reveal their human orthologoues. To handle the large volume of data that the RPE cDNA library project has generated a highly efficient and user-friendly RDBMS was designed. Using RDBMS data storage can be managed efficiently and flexibly. The RDBMS allows displaying the results in query-based form and report format with additional annotations, links and search functions. Out of the 341 known and predicted genes identified in this study, 2 were further analyzed. The RPE or/and retina specificity of these two clones were further confirmed by RT-PCR analysis in adult human tissues. Construction of a single nucleotide polymphism (SNP) map was initiated as a first step in future case/control association studies. SNP genotyping was carried out for one of these two clones (RPE01-D2, now known as RDH12). 12 SNPs were identified from direct sequencing of the 23.4-kb region, of which 5 are of high frequency. In a next step, comparison of allele frequencies between AMD patients and healthy controls is required. Completion of the expression analysis for other predicted genes identified during this study is in progress using real time RT-PCR and will provide additional candidate genes for further analyses. This study is expected to contribute to our understanding of the genetic basis of RPE function and to clarify the role of the RPE-expressed genes in the predisposition to AMD. It may also help reveal the mechanisms and pathways that are involved in the development of AMD or other retinal dystrophies.},
  subject      = {Senile Makuladegeneration},
  language  = {en}
}