Completed Theses 2013

Protein intrinsic disorder in the human herpesviruses

The discovery of intrinsically disordered proteins challenged the traditional protein structure paradigm, which stated that protein function derives from a speciVc, well-deVned structure. Protein disorder in the Herpesviridae was the object of the study, using four viruses as representatives and the mimivirus (APMV) as a comparison, as well as some model organisms (D. melanogaster, S. cerevisiae, C. elegans, H. sapiens, A. thaliansis). The overall disorder content was analyzed using various metrics. It was found that the Herpesviridae show an exceptionally high disorder content. Further analysis was concerned with the conservation of disorder across the herpesviruses. In a set of 30 genes conserved across all four herpesviruses of the study 14 had conserved disorder patterns. In genes less conserved the disorder diUerentiated between diUerent herpesvirus families and species. Using available protein-protein interaction data for the proteins of the conserved cluster showed that proteins that interact with host proteins tend to also have conserved disorder, while proteins responsible for interactions with viral proteins tend to have less conserved disorder, or none at all. Future investigation could be focused on determining whether the conserved disorder is important for the interactions or not. The incompleteness and uneven quality of currently available data clearly present a challenge to the analysis. The availability of more, better data will allow the expansion of the results of this study.

Bachelor thesis
Student: Nikolaos Papadopoulos
Supervisor: Burkhard Rost, Arthur Dong

Active Site Prediction

The pediction of active sites, catayalytic residues and structure stabilizing protein protein interactions is a common problem usually approached by homology modelling or profile profile alignments, mapping known sites to the target. In this thesis, we evaluate whether a sequence-only based method using features predicted by Predict Protein and SNAP can find such residues, too. This would allow us to annotate sequences with with virtually no identity to known structures. A good filter for active site candidates could also reducing the wet-lab effort to determine new catalytic sites.

Bachelor thesis
Student: Manfred Roos
Supervisor: Burkhard Rost, Andrea Schafferhans

Improvement of sequence based protein function prediction

Determining the function of a protein in biology is a very difficult, time consuming and expensive procedure. The most common vocabulary for defining different protein functions is the Gene Ontology, which defines protein functions in the categories: Cellular Component, Molecular function and Biological process. A lot of experiments are necessary to assign a protein correctly to these independent categories. Computational tools that are able to predict which function is most probable for a certain protein, and which can perhaps be excluded, save time and money. A recent competition, which aims to gather new function-prediction methods and test their performance and accuracy, has shown that there is a deficit, looking at the currently available prediction methods. Although there are various protein function prediction methods out there, the reliability of those is not satisfying. The approach presented in this thesis focuses on a method based on the prediction of protein function from its sequence alone. The target is to develop an automated function prediction method that outperforms former methods.

Diploma thesis
Student: Michael Kiening
Supervisor: Burkhard Rost, Tobias Hamp

Understanding molecular mechanism: analysis of the SLAC1 protein family

In 2010, the high-resolution structure of TehA was solved. TehA is a transmembrane protein with 10 helices, an anion channel and a homolog to SLAC1. SLAC1 is involved in the regulation of stomata opening and closing in leaves. In this study, we are researching the SLAC1 protein family to gain insight in the molecular mechanism of the proteins using implications from sequence homology. Moreover, electrostatic calculations and molecular dynamics simulations of the TehA structure are carried out.

Master thesis
Student: Eva Reisinger
Supervisor: Burkhard Rost, Edda Kloppmann