The MHC genomic region in most species is extremely polymorphic. The distinct specificity of the majority of the MHC molecules remains uncharacterized. The MHCcluster is a tool to functionally cluster MHC class I molecules (MHCI) based on their predicted binding specificity. The tool provides highly intuitive heat-map and graphical tree-based visualizations of the functional relationship between MHC class I and class II variants.
For MHC class I, peptide binding predictions are made using
NetMHCpan
PMID 19002680.
For MHC class II, peptide binding predictions are made using
NetMHCIIpan
PMID 21073747.
Note, as of July 8, 2014 peptide binding predictions can be made using either NetMHCpan version 2.8, or NetMHCpan version 2.7 (the earlier default)