@article{VainshteinSanchezBrazmaetal.2010, author = {Vainshtein, Yevhen and Sanchez, Mayka and Brazma, Alvis and Hentze, Matthias W. and Dandekar, Thomas and Muckenthaler, Martina U.}, title = {The IronChip evaluation package: a package of perl modules for robust analysis of custom microarrays}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-67869}, year = {2010}, abstract = {Background: Gene expression studies greatly contribute to our understanding of complex relationships in gene regulatory networks. However, the complexity of array design, production and manipulations are limiting factors, affecting data quality. The use of customized DNA microarrays improves overall data quality in many situations, however, only if for these specifically designed microarrays analysis tools are available. Results: The IronChip Evaluation Package (ICEP) is a collection of Perl utilities and an easy to use data evaluation pipeline for the analysis of microarray data with a focus on data quality of custom-designed microarrays. The package has been developed for the statistical and bioinformatical analysis of the custom cDNA microarray IronChip but can be easily adapted for other cDNA or oligonucleotide-based designed microarray platforms. ICEP uses decision tree-based algorithms to assign quality flags and performs robust analysis based on chip design properties regarding multiple repetitions, ratio cut-off, background and negative controls. Conclusions: ICEP is a stand-alone Windows application to obtain optimal data quality from custom-designed microarrays and is freely available here (see "Additional Files" section) and at: http://www.alice-dsl.net/evgeniy. vainshtein/ICEP/}, subject = {Microarray}, language = {en} } @phdthesis{Vainshtein2010, author = {Vainshtein, Yevhen}, title = {Applying microarray-based techniques to study gene expression patterns: a bio-computational approach}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-51967}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2010}, abstract = {The regulation and maintenance of iron homeostasis is critical to human health. As a constituent of hemoglobin, iron is essential for oxygen transport and significant iron deficiency leads to anemia. Eukaryotic cells require iron for survival and proliferation. Iron is part of hemoproteins, iron-sulfur (Fe-S) proteins, and other proteins with functional groups that require iron as a cofactor. At the cellular level, iron uptake, utilization, storage, and export are regulated at different molecular levels (transcriptional, mRNA stability, translational, and posttranslational). Iron regulatory proteins (IRPs) 1 and 2 post-transcriptionally control mammalian iron homeostasis by binding to iron-responsive elements (IREs), conserved RNA stem-loop structures located in the 5'- or 3'- untranslated regions of genes involved in iron metabolism (e.g. FTH1, FTL, and TFRC). To identify novel IRE-containing mRNAs, we integrated biochemical, biocomputational, and microarray-based experimental approaches. Gene expression studies greatly contribute to our understanding of complex relationships in gene regulatory networks. However, the complexity of array design, production and manipulations are limiting factors, affecting data quality. The use of customized DNA microarrays improves overall data quality in many situations, however, only if for these specifically designed microarrays analysis tools are available. Methods In this project response to the iron treatment was examined under different conditions using bioinformatical methods. This would improve our understanding of an iron regulatory network. For these purposes we used microarray gene expression data. To identify novel IRE-containing mRNAs biochemical, biocomputational, and microarray-based experimental approaches were integrated. IRP/IRE messenger ribonucleoproteins were immunoselected and their mRNA composition was analysed using an IronChip microarray enriched for genes predicted computationally to contain IRE-like motifs. Analysis of IronChip microarray data requires specialized tool which can use all advantages of a customized microarray platform. Novel decision-tree based algorithm was implemented using Perl in IronChip Evaluation Package (ICEP). Results IRE-like motifs were identified from genomic nucleic acid databases by an algorithm combining primary nucleic acid sequence and RNA structural criteria. Depending on the choice of constraining criteria, such computational screens tend to generate a large number of false positives. To refine the search and reduce the number of false positive hits, additional constraints were introduced. The refined screen yielded 15 IRE-like motifs. A second approach made use of a reported list of 230 IRE-like sequences obtained from screening UTR databases. We selected 6 out of these 230 entries based on the ability of the lower IRE stem to form at least 6 out of 7 bp. Corresponding ESTs were spotted onto the human or mouse versions of the IronChip and the results were analysed using ICEP. Our data show that the immunoselection/microarray strategy is a feasible approach for screening bioinformatically predicted IRE genes and the detection of novel IRE-containing mRNAs. In addition, we identified a novel IRE-containing gene CDC14A (Sanchez M, et al. 2006). The IronChip Evaluation Package (ICEP) is a collection of Perl utilities and an easy to use data evaluation pipeline for the analysis of microarray data with a focus on data quality of custom-designed microarrays. The package has been developed for the statistical and bioinformatical analysis of the custom cDNA microarray IronChip, but can be easily adapted for other cDNA or oligonucleotide-based designed microarray platforms. ICEP uses decision tree-based algorithms to assign quality flags and performs robust analysis based on chip design properties regarding multiple repetitions, ratio cut-off, background and negative controls (Vainshtein Y, et al., 2010).}, subject = {Microarray}, language = {en} } @phdthesis{Reinboth2012, author = {Reinboth, Jennifer}, title = {Cellular Factors Contributing to Host Cell Permissiveness in Support of Oncolytic Vaccinia Virus Replication}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-85392}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2012}, abstract = {In initial experiments, the well characterized VACV strain GLV-1h68 and three wild-type LIVP isolates were utilized to analyze gene expression in a pair of autologous human melanoma cell lines (888-MEL and 1936 MEL) after infection. Microarray analyses, followed by sequential statistical approaches, characterized human genes whose transcription is affected specifically by VACV infection. In accordance with the literature, those genes were involved in broad cellular functions, such as cell death, protein synthesis and folding, as well as DNA replication, recombination, and repair. In parallel to host gene expression, viral gene expression was evaluated with help of customized VACV array platforms to get better insight over the interplay between VACV and its host. Our main focus was to compare host and viral early events, since virus genome replication occurs early after infection. We observed that viral transcripts segregated in a characteristic time-specific pattern, consistent with the three temporal expression classes of VACV genes, including a group of genes which could be classified as early-stage genes. In this work, comparison of VACV early replication and respective early gene transcription led to the identification of seven viral genes whose expression correlated strictly with replication. We considered the early expression of those seven genes to be representative for VACV replication and we therefore referred to them as viral replication indicators (VRIs). To explore the relationship between host cell transcription and viral replication, we correlated viral (VRI) and human early gene expression. Correlation analysis revealed a subset of 114 human transcripts whose early expression tightly correlated with early VRI expression and thus early viral replication. These 114 human molecules represented an involvement in broad cellular functions. We found at least six out of 114 correlates to be involved in protein ubiquitination or proteasomal function. Another molecule of interest was the serine-threonine protein kinase WNK lysine-deficient protein kinase 1 (WNK1). We discovered that WNK1 features differences on several molecular biological levels associated with permissiveness to VACV infection. In addition to that, a set of human genes was identified with possible predictive value for viral replication in an independent dataset. A further objective of this work was to explore baseline molecular biological variances associated with permissiveness which could help identifying cellular components that contribute to the formation of a permissive phenotype. Therefore, in a subsequent approach, we screened a set of 15 melanoma cell lines (15-MEL) regarding their permissiveness to GLV-1h68, evaluated by GFP expression levels, and classified the top four and lowest four cell lines into high and low permissive group, respectively. Baseline gene transcriptional data, comparing low and highly permissive group, suggest that differences between the two groups are at least in part due to variances in global cellular functions, such as cell cycle, cell growth and proliferation, as well as cell death and survival. We also observed differences in the ubiquitination pathway, which is consistent with our previous results and underlines the importance of this pathway in VACV replication and permissiveness. Moreover, baseline microRNA (miRNA) expression between low and highly permissive group was considered to provide valuable information regarding virus-host co-existence. In our data set, we identified six miRNAs that featured varying baseline expression between low and highly permissive group. Finally, copy number variations (CNVs) between low and highly permissive group were evaluated. In this study, when investigating differences in the chromosomal aberration patterns between low and highly permissive group, we observed frequent segmental amplifications within the low permissive group, whereas the same regions were mostly unchanged in the high group. Taken together, our results highlight a probable correlation between viral replication, early gene expression, and the respective host response and thus a possible involvement of human host factors in viral early replication. Furthermore, we revealed the importance of cellular baseline composition for permissiveness to VACV infection on different molecular biological levels, including mRNA expression, miRNA expression, as well as copy number variations. The characterization of human target genes that influence viral replication could help answering the question of host cell response to oncolytic virotherapy and provide important information for the development of novel recombinant vaccinia viruses with improved features to enhance replication rate and hence trigger therapeutic outcome.}, subject = {Vaccinia-Virus}, language = {en} } @phdthesis{Friedrich2009, author = {Friedrich, Torben}, title = {New statistical Methods of Genome-Scale Data Analysis in Life Science - Applications to enterobacterial Diagnostics, Meta-Analysis of Arabidopsis thaliana Gene Expression and functional Sequence Annotation}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-39858}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2009}, abstract = {Recent progresses and developments in molecular biology provide a wealth of new but insufficiently characterised data. This fund comprises amongst others biological data of genomic DNA, protein sequences, 3-dimensional protein structures as well as profiles of gene expression. In the present work, this information is used to develop new methods for the characterisation and classification of organisms and whole groups of organisms as well as to enhance the automated gain and transfer of information. The first two presented approaches (chapters 4 und 5) focus on the medically and scientifically important enterobacteria. Its impact in medicine and molecular biology is founded in versatile mechanisms of infection, their fundamental function as a commensal inhabitant of the intestinal tract and their use as model organisms as they are easy to cultivate. Despite many studies on single pathogroups with clinical distinguishable pathologies, the genotypic factors that contribute to their diversity are still partially unknown. The comprehensive genome comparison described in Chapter 4 was conducted with numerous enterobacterial strains, which cover nearly the whole range of clinically relevant diversity. The genome comparison constitutes the basis of a characterisation of the enterobacterial gene pool, of a reconstruction of evolutionary processes and of comprehensive analysis of specific protein families in enterobacterial subgroups. Correspondence analysis, which is applied for the first time in this context, yields qualitative statements to bacterial subgroups and the respective, exclusively present protein families. Specific protein families were identified for the three major subgroups of enterobacteria namely the genera Yersinia and Salmonella as well as to the group of Shigella and E. coli by applying statistical tests. In conclusion, the genome comparison-based methods provide new starting points to infer specific genotypic traits of bacterial groups from the transfer of functional annotation. Due to the high medical importance of enterobacterial isolates their classification according to pathogenicity has been in focus of many studies. The microarray technology offers a fast, reproducible and standardisable means of bacterial typing and has been proved in bacterial diagnostics, risk assessment and surveillance. The design of the diagnostic microarray of enterobacteria described in chapter 5 is based on the availability of numerous enterobacterial genome sequences. A novel probe selection strategy based on the highly efficient algorithm of string search, which considers both coding and non-coding regions of genomic DNA, enhances pathogroup detection. This principle reduces the risk of incorrect typing due to restrictions to virulence-associated capture probes. Additional capture probes extend the spectrum of applications of the microarray to simultaneous diagnostic or surveillance of antimicrobial resistance. Comprehensive test hybridisations largely confirm the reliability of the selected capture probes and its ability to robustly classify enterobacterial strains according to pathogenicity. Moreover, the tests constitute the basis of the training of a regression model for the classification of pathogroups and hybridised amounts of DNA. The regression model features a continuous learning capacity leading to an enhancement of the prediction accuracy in the process of its application. A fraction of the capture probes represents intergenic DNA and hence confirms the relevance of the underlying strategy. Interestingly, a large part of the capture probes represents poorly annotated genes suggesting the existence of yet unconsidered factors with importance to the formation of respective virulence phenotypes. Another major field of microarray applications is gene expression analysis. The size of gene expression databases rapidly increased in recent years. Although they provide a wealth of expression data, it remains challenging to integrate results from different studies. In chapter 6 the methodology of an unsupervised meta-analysis of genome-wide A. thaliana gene expression data sets is presented, which yields novel insights in function and regulation of genes. The application of kernel-based principal component analysis in combination with hierarchical clustering identified three major groups of contrasts each sharing overlapping expression profiles. Genes associated with two groups are known to play important roles in Indol-3 acetic acid (IAA) mediated plant growth and development as well as in pathogen defence. Yet uncharacterised serine-threonine kinases could be assigned to novel functions in pathogen defence by meta-analysis. In general, hidden interrelation between genes regulated under different conditions could be unravelled by the described approach. HMMs are applied to the functional characterisation of proteins or the detection of genes in genome sequences. Although HMMs are technically mature and widely applied in computational biology, I demonstrate the methodical optimisation with respect to the modelling accuracy on biological data with various distributions of sequence lengths. The subunits of these models, the states, are associated with a certain holding time being the link to length distributions of represented sequences. An adaptation of simple HMM topologies to bell-shaped length distributions described in chapter 7 was achieved by serial chain-linking of single states, while residing in the class of conventional HMMs. The impact of an optimisation of HMM topologies was underlined by performance evaluations with differently adjusted HMM topologies. In summary, a general methodology was introduced to improve the modelling behaviour of HMMs by topological optimisation with maximum likelihood and a fast and easily implementable moment estimator. Chapter 8 describes the application of HMMs to the prediction of interaction sites in protein domains. As previously demonstrated, these sites are not trivial to predict because of varying degree in conservation of their location and type within the domain family. The prediction of interaction sites in protein domains is achieved by a newly defined HMM topology, which incorporates both sequence and structure information. Posterior decoding is applied to the prediction of interaction sites providing additional information of the probability of an interaction for all sequence positions. The implementation of interaction profile HMMs (ipHMMs) is based on the well established profile HMMs and inherits its known efficiency and sensitivity. The large-scale prediction of interaction sites by ipHMMs explained protein dysfunctions caused by mutations that are associated to inheritable diseases like different types of cancer or muscular dystrophy. As already demonstrated by profile HMMs, the ipHMMs are suitable for large-scale applications. Overall, the HMM-based method enhances the prediction quality of interaction sites and improves the understanding of the molecular background of inheritable diseases. With respect to current and future requirements I provide large-scale solutions for the characterisation of biological data in this work. All described methods feature a highly portable character, which allows for the transfer to related topics or organisms, respectively. Special emphasis was put on the knowledge transfer facilitated by a steadily increasing wealth of biological information. The applied and developed statistical methods largely provide learning capacities and hence benefit from the gain of knowledge resulting in increased prediction accuracies and reliability.}, subject = {Genomik}, language = {en} } @phdthesis{Engelmann2008, author = {Engelmann, Julia Cath{\´e}rine}, title = {DNA microarrays: applications and novel approaches for analysis and interpretation}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-29747}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2008}, abstract = {In der vorliegenden Dissertation wird die Entwicklung eines phylogenetischen DNA Microarrays, die Analyse von mehreren Microarray-Genexpressionsdatens{\"a}tzen und neue Ans{\"a}tze f{\"u}r die Datenanalyse und Interpretation der Ergebnisse vorgestellt. Die Entwicklung und Analyse der Daten eines phylogenetischen DNA Microarrays wird in der ersten Publikation dargestellt. Ich konnte zeigen, dass die Spezies-Detektion mit phylogenetischen Microarrays durch die Datenanalyse mit einem linearen Regressionsansatz signifikant verbessert werden kann. Standard-Methoden haben bislang nur Signalintensit{\"a}ten betrachtet und eine Spezies als an- oder abwesend bezeichnet, wenn die Signalintensit{\"a}t ihres Messpunktes oberhalb eines willk{\"u}rlich gesetzten Schwellenwertes lag. Dieses Verfahren ist allerdings aufgrund von Kreuz-Hybridisierungen nicht auf sehr nah verwandte Spezies mit hoher Sequenzidentit{\"a}t anwendbar. Durch die Modellierung des Hybridisierungs und Kreuz-Hybridisierungsverhaltens mit einem linearen Regressionsmodell konnte ich zeigen, dass Spezies mit einer Sequenz{\"a}hnlichkeit von 97\% im Markergen immer noch unterschieden werden k{\"o}nnen. Ein weiterer Vorteil der Modellierung ist, dass auch Mischungen verschiedener Spezies zuverl{\"a}ssig vorhergesagt werden k{\"o}nnen. Theoretisch sind auch quantitative Vorhersagen mit diesem Modell m{\"o}glich. Um die großen Datenmengen, die in {\"o}ffentlichen Microarray-Datenbanken abgelegt sind besser nutzen zu k{\"o}nnen, bieten sich Meta-Analysen an. In der zweiten Publikation wird eine explorative Meta-Analyse auf Arabidopsis thaliana-Datens{\"a}tzen vorgestellt. Mit der Analyse verschiedener Datens{\"a}tze, die den Einfluss von Pflanzenhormonen, Pathogenen oder verschiedenen Mutationen auf die Genexpression untersucht haben, konnten die Datens{\"a}tze anhand ihrer Genexpressionsprofile in drei große Gruppen eingeordnet werden: Experimente mit Indol-3-Essigs{\"a}ure (IAA), mit Pathogenen und andere Experimente. Gene, die charakteristisch f{\"u}r die Gruppe der IAA-Datens{\"a}tze beziehungsweise f{\"u}r die Gruppe der Pathogen-Datens{\"a}tze sind, wurden n{\"a}her betrachtet. Diese Gene hatten Funktionen, die bereits mit Pathogenbefall bzw. dem Einfluss von IAA in Verbindung gebracht wurden. Außerdem wurden Hypothesen {\"u}ber die Funktionen von bislang nicht annotierten Genen aufgestellt. In dieser Arbeit werden auch Prim{\"a}ranalysen von einzelnen Arabidopsis thaliana Genexpressions-Datens{\"a}tzen vorgestellt. In der dritten Publikation wird ein Experiment beschrieben, das durchgef{\"u}hrt wurde um herauszufinden ob Mikrowellen-Strahlung einen Einfluss auf die Genexpression einer Zellkultur hat. Dazu wurden explorative Analysemethoden angewendet. Es wurden geringe aber signifikante Ver{\"a}nderungen in einer sehr kleinen Anzahl von Genen beobachtet, die experimentell best{\"a}tigt werden konnten. Die Funktionen der regulierten Gene und eine Meta-Analyse mit {\"o}ffentlich zug{\"a}nglichen Datens{\"a}tzen einer Datenbank deuten darauf hin, dass die pflanzliche Zellkultur die Strahlung als eine Art Energiequelle {\"a}hnlich dem Licht wahrnimmt. Des weiteren wird in der vierten Publikation die funktionelle Analyse eines Arabidopsis thaliana Genexpressionsdatensatzes beschrieben. Die Analyse der Genexpressions eines pflanzlichen Tumores zeigte, dass er seinen Stoffwechsel von aerob und auxotroph auf anaerob und heterotroph umstellt. Gene der Photosynthese werden im Tumorgewebe reprimiert, Gene des Aminos{\"a}ure- und Fettstoffwechsels, der Zellwand und Transportkan{\"a}le werden so reguliert, dass Wachstum und Entwicklung des Tumors gef{\"o}rdert werden. In der f{\"u}nften Publikation in dieser Arbeit wird GEPAT (Genome Expression Pathway Analysis Tool) beschrieben. Es besteht aus einer Internet- Anwendung und einer Datenbank, die das einfache Hochladen von Datens{\"a}tzen in die Datenbank und viele M{\"o}glichkeiten der Datenanalyse und die Integration anderer Datentypen erlaubt. In den folgenden zwei Publikationen (Publikation 6 und Publikation 7) wird GEPAT auf humane Microarray-Datens{\"a}tze angewendet um Genexpressionsdaten mit weiteren Datentypen zu verkn{\"u}pfen. Genexpressionsdaten und Daten aus vergleichender Genom-Hybridisierung (CGH) von prim{\"a}ren Tumoren von 71 Mantel-Zell-Lymphom (MCL) Patienten erm{\"o}glichte die Ermittlung eines Pr{\"a}diktors, der die Vorhersage der {\"U}berlebensdauer von Patienten gegen{\"u}ber herk{\"o}mmlichen Methoden verbessert. Die Analyse der CGH Daten zeigte, dass auch diese f{\"u}r die Vorhersage der {\"U}berlebensdauer geeignet sind. F{\"u}r den Datensatz von Patienten mit großzellig diffusem B-Zell-Lymphom DLBCL konnte aus den Genexpressionsdaten ebenfalls ein neuer Pr{\"a}diktor vorgeschlagen werden. Mit den zwischen lang und kurz {\"u}berlebenden Patienten differentiell exprimierten Genen der MCL Patienten und mit den Genen, die zwischen den beiden Untergruppen von DLBCL reguliert sind, wurden Interaktionsnetzwerke gebildet. Diese zeigen, dass bei beiden Krebstypen Gene des Zellzyklus und der Proliferation zwischen Patienten mit kurzer und langer {\"U}berlebensdauer unterschiedlich reguliert sind.}, subject = {Microarray}, language = {en} } @phdthesis{Busold2006, author = {Busold, Christian}, title = {Facilitating functional interpretation of microarray data by integration of gene annotations in Correspondence Analysis}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-21150}, school = {Universit{\"a}t W{\"u}rzburg}, year = {2006}, abstract = {DNS-Chips ('Microarrays') haben sich zu einer der Standardmethoden zur Erstellung von genomweiten Expressionsstudien entwickelt. Mittlerweile wurden dazu eine Vielzahl von Methoden zur Identifizierung von differentiell regulierten Genen ver{\"o}ffentlicht. Ungeachtet dessen stellt die abschliessende funktionelle Interpretation der Ergebnisse einen der Engp{\"a}sse in der Analyse von Chip-Daten dar. Die Mehrzahl der Analysemethoden stellt die signifikant regulierten Gene in Listen dar, aus denen in einem weiteren Schritt gemeinsame funktionelle Eigenschaften abgeleitet werden m{\"u}ssen. Dies stellt nicht nur eine arbeitsintensive Arbeit dar, die mit steigender Anzahl an experimentellen Konditionen immer weniger praktikabel wird, sondern ist auch fehleranf{\"a}llig, da diese Auswertung im allgemeinen auf dem visuellen Vergleich von Listen beruht. In der vorliegenden Arbeit wurden Methoden f{\"u}r eine rechnergest{\"u}tzte Auswertung von funktionellen Geneigenschaften entwickelt und validiert. Hierzu wurde die 'Gene Ontology' als Quelle f{\"u}r die Annotationsdaten ausgew{\"a}hlt, da hier die Daten in einem Format gespeichert sind, das sowohl eine leichte menschliche Interaktion sowie die statistische Analyse der Annotationen erm{\"o}glicht. Diese Genannotation wurden als Zusatzinformationen in die Korrespondenzanalyse integriert, welches eine simultane Darstellung von Genen, Hybridisierungen und funktionellen Kategorien in einer Grafik erm{\"o}glicht. Aufgrund der st{\"a}ndig wachsenden Anzahl an verf{\"u}gbaren Annotationen und der Tatsache, daß zwischen den meisten experimentellen Bedingungen nur wenige funktionelle Prozesse differentiell reguliert sind, wurden Filter entwickelt, die die Anzahl der dargestellten Annotationen auf eine im gegebenen experimentellen Kontext relevante Gruppe reduzieren. Die Anwendbarkeit der Visualisierung und der Filter wurde auf Datens{\"a}tzen unterschiedlicher Komplexit{\"a}t getestet: beginnend mit dem gut verstandenen Glukosestoffwechsel im Modellorganismus S. cerevisiae, bis hin zum Vergleich unterschiedlicher Tumortypen im Menschen. In beiden F{\"a}llen generierte die Methode gut zu interpretierende Grafiken, in denen die funktionellen Hauptunterschiede durch die dargestellten Annotationen gut beschrieben werden [90]. W{\"a}hrend die Integration von Annotationsdaten wie GO die funktionelle Interpretation vereinfacht, fehlt die M{\"o}glichkeit zur Identifikation einzelner relevanter Schl{\"u}sselgene. Um eine solche Analyse zu erm{\"o}glichen, wurden Daten zum Vorkommen von Transskriptionsfaktorbindestellen in den 5'-Bereichen von Genen integriert. Auch diese Methode wurde an Datens{\"a}tzen von S. cerevisiae und vergleichenden Studien von humanen Krebszelllinien validiert.In beiden F{\"a}llen konnten Transkriptionsfaktoren identifiziert werden, die f{\"u}r die beobachteten transkriptionellen Unterschiede von entscheidender Bedeutung sind [206]. Zusammenfassend, erm{\"o}glicht die Integration von Zusatzinformationen in die Korrespondenzanalyse eine simultane Visualisierung von Genen, Hybridisierungen und Annotationsdaten in einer einzigen, gut zu interpretierenden Grafik. Dies erlaubt auch in komplexen experimentellen Bedingungen eine intuitive Identifizierung von relevanten Annotationen. Der hier vorgestellte Ansatz, ist nicht auf die gezeigten Datenstrukturen beschr{\"a}nkt, sondern kann auf die Mehrzahl der verf{\"u}gbaren Annotationsdaten angewendet werden.}, subject = {Microarray}, language = {en} }