Functional Sites in Structure and Sequence. Protein Active Sites and miRNA Target Recognition -

The number of protein three-dimensional structures is increasing steeply, and structural genomics projects aim to solve the structures for all proteins as a means to understanding function. In the first part of my thesis, I developed a method for the comparison of local structural patterns (e.g. enzyme active sites) that provides a reliable statistical measure to discern meaningful matches from noise. The method is complementary to structural alignment as it is able to confirm functional similarities suggested by an overall similar structure but also detects functional similarities between different folds. An easy-to-use interface is available on the Internet for functional annotation of protein structures (http://pints.embl.de). In the second part of my thesis, I present a computational screen for microRNA (miRNA) targets in Drosophila. miRNAs are short RNAs that inhibit translation of target messenger RNAs in animals by binding to complementary sites in their 3� untranslated regions. Target predictions were urgently needed as targets were known for only three of the more than 700 miRNAs. Of my predictions, six were validated experimentally and others are likely to be functional, making the results a useful resource for miRNA research. The screen extended miRNA function to pathway control, nervous system development and regulation of metabolism, and revealed that one miRNA typically regulates several targets but also that one gene is likely to be targeted by several miRNAs