Preliminary Trial Version
SVMTurn is a method for the prediction of regular turn regions in proteins. The predictions are based on
information encoded in the amino acid sequence. A jury of trained support vector machine (SVM) models decides which parts
of a sequence can be classified as turns. This decision depends on a threshold value. Two different threshold values can be chosen by the user,
the default value has been empirically chosen and gives best performance with a small rate of false positives, a threshold value
of 0 is more lush and leads to more results - but possibly to more false positives as well.
Predictions are available for turns of length 4-6 residues (for definitions see reference below) and may overlap each other.
More information can be found here:
Meissner M., Koch O., Klebe G., Schneider G. (2008) Prediction of turn types in protein structure by machine-learning classifiers, Proteins 74, 344-352.
Please cite this reference if you publish results obtained via this webserver.