SVMHC information page
Input formats for sequences
-
One letter codes for the protein sequences should be used.
(i.e ACDEFGHIKLMNPQRSTVWY)
Please note that the sequence should be in CAPITAL LETTERS!!!
No other formats are accepted
Output formats
- The output from the prediction is a ranked list of predicted binders (scores of 0 or higher). The
sequence, score and position are given for each predicted binder. It also possible to download ALL the results from the prediction. The
format is a ranked list of both predicted binders and non-binders (scores below 0). The typical range of ouputs from the server is -1 to +1, where 0 is the default cutoff used to discriminate binders from non-binders.
- The predictions are also presented in a way that highlight regions with a
lot of binders in a protein. An amino acid predicted to be in a binding
peptide is highlighted in red and other residues are in black.
Support Vector Machines
The SVM package used for prediction is SVMlight by Thorsten Joachims. The source code is free for scientific use and can be downloaded from http://www.cs.cornell.edu/People/tj/svm_light/.
A comprehensive introduction to SVMs is given in An introduction to Support Vector Machines and other kernel-based learning methods by Nello Cristianini and John Shawe-Taylor.
How to cite the prediction service
Prediction of
MHC class I binding peptides, using SVMHC. Pierre Dönnes and Arne Elofsson
in: BMC
Bioinformatics 2002 3: 25
For further information about SVMHC, please contactPierre
Dönnes or Arne Elofsson.