CAMPS

Content:


Why does CAMPS only cover α-helical membrane proteins and not β-barrels as well?
Because the prediction of β-barrels is more difficult and they are less abundant (only 2-3%) than α-helical membrane proteins.

[Top]


Why do you use sequence clustering for structure classification? Almost all traditional approaches for protein structure classification (such as SCOP and CATH) perform structure-structure comparisons and use the structure similarity scores for clustering.
Of course, this is one possibility to do structural classification for membrane proteins. But the drawback of this method is that due to the paucity of structural data (only 2% of all structures in the PDB account for membrane proteins!) such a classification would cover only a small portion of the membrane protein space. That's why we decided to do the clustering at the sequence level in order to provide a comprehensive membrane protein structure classification.

[Top]


What is the meaning of the different classification levels?
The CAMPS database provides a hierarchical clustering of α-helical membrane proteins. The SC-clusters (where SC stands for structurally correlated) constitute the first layer of the whole classification. Members of a SC-cluster are similar in sequence and have similar transmembrane helix numbers and loop length patterns. They are likely to share the same fold.
The FH-clusters (functionally homogeneous) are subclusters of the SC-clusters containing membrane proteins with the same or similar domain architecture. MD-clusters (modeling distance) are also subclusters of SC-clusters and their members share a sequence identity of at least 30%. FH-clusters and MD-clusters together form the second layer of the classification.
Thus, the SC-,FH, and MD-clusters approach different levels: fold, function and modeling distance.

[Top]

Technische Universität München - Department of Genome Oriented Bioinformatics