Computational proteomics: from numbers to biology

Abstract

The field of biological research is increasingly trying to understand the complex processes that are too small to observe by eye, even when using a microscope. The real action is taking place at the molecular level; therefore, the only way to investigate this is to use molecular tools. Modern techniques allow studying all proteins in a sample, and the prevailing tool is mass spectrometry based proteomics. As a high-throughput technique, computational analysis is indispensable. In this thesis, we describe several aspects of the treatment of proteomics data using in-house developed tools, and the analysis of several biological experiments. In a first example to illustrate how we used mass spectrometry to investigate biological problems, we investigated the evolution of tyrosine phosphorylation, which possibly arose simultaneously with multicellularity. To investigate this, the placozoan Trichoplax adhaerens was analysed, which is arguably the most primitive animal. We observed several tyrosine kinases in the genome sequence, and phosphoproteomics analysis revealed an extraordinary amount of tyrosine phosphorylation events compared to modern animals. In a second example, we simultaneously analysed RNA, DNA and proteins in a study of a single organ, the rat liver. We compared two different rat strains, the Brown Norwegian (BN) rat, and a variant exhibiting spontaneous hypertension (SHR). Through these measurements we found a differentially regulated rat homolog of a gene that is linked to hypertension in human, CYP17A1, both at the RNA and at the proteome level. The DNA sequence showed a polymorphism in the gene promoter, indicating that a single point mutation may be responsible for the mis-regulation of this gene and protein.