|
|||||||||||||||
 |
|
||||||||||||||
|
|||||||||||||||
| |||||||||||||||
|
Protein post-translational modification (PTM)
Group leader: Henrik Nielsen CBS has a long history of predicting various types of protein PTMs (post-translational modifications). Among the first publically available CBS prediction servers was SignalP, which predicts secretory signal peptides and their cleavage sites. Since the launch in 1996, it has been heavily used by the molecular biology community, and it is still among our most popular tools. It should be noted that although the "P" in "PTM" stands for "post-", some of these modifications actually occur during translation. The recognition and cleavage of signal peptides, for example, is predominantly co-translational in eukaryotes, but post-translational in bacteria. Most PTMs occur at specific, yet variable, motifs in the target proteins. In many cases, the motifs are too variable to be recognized by simple consensus patterns. Instead, machine learning techniques, such as artificial neural networks (NNs) or hidden Markov models (HMMs), are often well suited to integrate the subtleties of local sequence context and global protein properties. In this context, information about secondary or teritiary structure (whether experimentally determined or computationally predicted) is often integrated into the input features. In most cases, the modifying enzymes must recognize the native three-dimensional structure around an acceptor site, and therefore structural motifs may be more conserved than sequence motifs. PTMs can roughly be divided into three types, mentioned below with examples of modification types for which the PTM group has finished or on-going projects:
Prediction of protein PTMs is an important research tool, not only for studying the modifications as such, but also for larger-scale systems biology studies, e.g. protein function prediction. There is still a lot of genes and gene products hidden in the "midnight zone" - i.e. orphan proteins with no detectable homology to any other protein of known function. Many projects are aimed at elucidating the functions of these orphan proteins, some by experimental approaches such as gene expression profiling/microarrays, others by predictive approaches. In the latter case, PTMs have been shown to be important features for determining protein function. This knowledge is utilized in an integrative systems biology approach involving both the PTM group and other CBS research groups, which has until now resulted in the prediction servers ProtFun and ArchaeaFun. The PTM group has on-going collaborations with several other CBS research groups, but our closest partner is the Metagenomics group. PTM and Metagenomics hold joint group meetings under the heading "PBI", which used to stand for "Protein bioinformatics". Selected Publications:
|
