Podcasts about gene ontology

The aim of the present work was to identify the genes that played a role in ecological adaptation in D. melanogaster. This species, which originated in Africa, successfully adapted to a broad range of climates during the last 100.000 years. To find the genes involved, I used two different approaches: (1) a genomic region containing several ecologically relevant candidate genes and putatively carrying footprints of selection was investigated using selective sweep mapping, and (2) cold tolerance that might have been an important phenotype for the adaptation to the temperate climates was investigated using a QTL analysis. Using the technique of selective sweep mapping pioneered in the Stephan’s group, I detected evidence that recent strong positive selection has been acting on a small DNA region of 2.7 kb overlapping with the 3’ end of the HDAC6 gene in the ancestral African population. This gene codes for a newly characterized cell stress surveillance factor. HDAC6 is an unusual histone-deacetylase. It is localized in the cytoplasm and has a ubiquitin-binding and a tubulin-deacetylase activity. These properties make HDAC6 a key regulator of cytotoxic stress resistance. The phenotypic analyses show that the African and the European populations have very strong cold tolerance differences. By removing the effects of the autosomes, I showed that a significant amount of the phenotypic variance is due to genetic factors carried by the X chromosome. These factors were then more precisely mapped to two genomic regions of the X chromosome. By comparing the present results with other association studies and the Gene Ontology database, it was possible to determine a list of candidate genes influencing cold tolerance in D. melanogaster. As this list is limited to a very small number of genes, additional investigations for footprints of selection in these regions may be used to confirm their role in ecological adaptation.

Die Analyse von Genexpressionsdaten, die durch die Microarray-Technologie bereit gestellt werden, ist in den letzten Jahren zu einem interessanten Forschungsfeld der Statistik geworden. Die ersten Verfahren auf diesem Gebiet zielen darauf ab, differentiell exprimierte Gene aus der riesigen Menge aller Gene eines Microarrays heraus zu filtern. Das Resultat einer solchen genweisen Analyse ist eine Liste interessanter Gene. Derartige Listen einzeln ausgewählter Gene sind allerdings schwer in einen biologischen Kontext zu bringen. Überdies hängen sie stark von der verwendeten Analysemethode und vom jeweiligen Datensatz ab, so daß Genlisten verschiedener Arbeitsgruppen meist eine relativ schlechte Übereinstimmung aufweisen. Eine Alternative beziehungsweise Weiterführung der genweisen Herangehensweise bietet die Analyse funktioneller Gengruppen. Diese beinhalten biologisches Vorwissen über das Zusammenspiel von Genen. Somit sind relevante Gengruppen sinnvoller interpretierbar als einzelne relevante Gene. Es werden verschiedene Verfahren für die Untersuchung funktioneller Gengruppen hinsichtlich differentieller Expression vorgestellt und auf methodischer Ebene sowie anhand von realen Datenbeispielen und Simulationsstudien verglichen. Von speziellem Interesse ist hier die Familie von Gengruppen, die durch die Gene Ontology definiert wird. Die hierarchische Struktur dieser Ontologien bedeutet eine zusätzliche Herausforderung für die Analyse, insbesondere für die Adjustierung für multiples Testen. Ein globaler Test auf differentielle Expression in Gengruppen ist das GlobalAncova Verfahren, welches im Rahmen dieser Arbeit weiter entwickelt und als R Paket bereit gestellt wurde. Die Signifikanz von Gengruppen kann dabei durch ein Permutationsmodell sowie über die asymptotische Verteilung der Teststatistik bewertet werden. Wir legen die theoretischen Grundlagen und Aspekte der Programmierung des Verfahrens dar. GlobalAncova eignet sich für die Analyse komplexer Fragestellungen. Hierzu werden einige ausführliche Auswertungen präsentiert, die im Rahmen von Kooperationen mit Medizinern und Biologen durchgeführt wurden.

This thesis applies quantitative mass spectrometry to research topics in relation to cancer. Proteome-wide quantification at the protein expression level and phosphorylation level were achieved. The technologies developed and used here cover the latest improvements in instrumentation in mass spectrometry, strategies in phosphopeptide enrichment in large scale, algorithms in data analysis and their streamlined implementation, and data mining in downstream bioinformatics. For each of the projects described in this thesis, proteome mapping routinely resulted in identification and quantitation of around 4,000 proteins and phosphoproteome mapping often lead to quantitation of more than 5,000 phosphorylation sites. This ‘systems-wide’ quantitation of the proteome and phosphoproteome is a completely novel development, which has not been used in cancer related topics before. Three major biology topics are studied in this thesis. In the first project, the phosphoproteome of a mouse liver cancer cell line Hepa1-6 was analyzed in-depth, by using phosphatase inhibitors (calyculin A, deltamethrin, and Na-pervanadate) to boost phosphorylation. The characterization of the phosphoproteome revealed a broad spectrum of cellular compartmentalization and biological functions. Quantitation of phosphatase inhibitor treatment using the Stable Isotope Labeling by Amino Acids in Cell culture (SILAC) method revealed the quantitative effects of these inhibitor compounds on the whole phosphoproteome. To our surprise, these three broadband phosphatase inhibitors displayed very different efficiency, with tyrosine phosphorylation significantly boosted but serine/threonine phosphorylation much less affected. Additionally, a method to estimate an upper bound of the stoichiometry of phosphorylation was introduced by comparing phosphorylation in three SILAC conditions: non-treated cells, stimulated cells (e.g. with insulin), and only phosphatase inhibitor treated cells. The methods developed here can be used directly in development of drugs directed against kinases and phosphatases, key regulators in cancer and other diseases. The second project continues with the application of phosphoproteomics techniques. Kinase inhibitors influence cellular signal transduction processes and therefore are of great potential in rescuing aberrant cellular signaling in tumors. In fact they constitute a significant portion of drug developing programs in pharmaceutical industry. With the aim of quantifying the effect of kinase inhibitors over the entire signaling network, the second project first set out to study two very commonly used kinase inhibitor compounds for MAPKs: U0126 and SB202190. Their effect on epidermal growth factor (EGF) signal transduction was quantified and compared using the HeLa cell system. The study confirmed that the MAPK cascades are the predominant signaling branches for propagating the EGF signaling at early time points of stimulation. These large scale examinations also suggest that U0126 and SB202190 are quite specific inhibitors for MAPKs as the majority of regulated phosphopeptides appears to belong to the MAPK pathways. In the second part of the project, the effect on phosphoproteome changes of the chemical compound dasatinib, which was demonstrated to effectively inhibit the constitutively activated fusion protein BCR-ABL and was recently approved for chronic myelogenous leukemia (CML) therapy, was quantified in the human CML cell line K562. Bioinformatic analysis revealed that the most influenced signal transduction branch was the Erk1/2 cascade. Overall more than 500 phosphorylation sites were found to be regulated by dasatinib, the vast majority not described in the literature yet. The third project compared the proteomes of mouse hepatoma cell line Hepa1-6 with the non-transformed mouse primary hepatocytes. This was performed by combining the SILAC heavy labeled form of Hepa1-6 with the primary hepatocytes. To characterize the features of these two proteomes, quantitation information (i.e. protein ratios between the two cell types) was used to divide all proteins into five quantiles. Each quantile was clustered according to the Gene Ontology and KEGG pathway databases to assess their enriched functional groups and signaling pathways. To integrate this information at a higher level, hierarchical clustering based on the p-value from the first Gene Ontology and KEGG clustering was performed. Using this improved bioinformatic algorithm for data mining, the proteomic phenotypes of the primary cells and transformed cells are immediately apparent. Primary hepatocytes are enriched in mitochondrial functions such as metabolic regulation and detoxification, as well as liver functions with tissue context such as secretion of plasma and low-density lipoprotein (LDL). In contrast, the transformed cancer cell line Hepa1-6 is enriched in cell cycle and growth functions. Interestingly, several aspects of the molecular basis of the “Warburg effect” described in many cancer cells became apparent in Hepa1-6, such as increased expression of glycolysis markers and decreased expression of markers for tricarboxylic acid (TCA) cycle. Studies in this thesis only provide examples of the application of mass spectrometry-based quantitative proteomics and phosphoproteomics in cancer research. The connection to clinical research, especially the assessment of drug effects on a proteome wide scale, is a specific feature of this thesis. Although this development is only in its infancy, it reflects a trend in the quantitative mass spectrometry field. We believe that more and more clinical related topics can and will be studied by these powerful methods.