Max-Planck-Institute for Infection Biology


Information and Hyperlinks

Charite : Institute for Biochemistry


MAPPP (MHC-I Antigenic Peptide Processing Prediction) will predict possible antigenic peptides to be processed and finally presented on the cell surfaces. It aides the prediction of immunodominant T-cell epitopes.
The processing and transporting is known to take multiple pathways:

Cytoplasmic protein Proteasomal processing alternative transport TAP transport MHC class I binding Peptide trimming Transport to cell surface Presentation to T cells T cell recognition Degradation Peptide presentation pathway
Cytoplasmic proteins are fragmented by the proteasomal processing.

The transport of the fragments to the ER by TAP or alternative transport mechanisms.

In the ER, further trimming of the peptides may occur.

If the peptides are correct sized, the bind to MHC class I molecules.

The MHC/peptide-complex takes its way to the cell surface.

Here, the antigenic peptide is presented to the T-cells.

If the antigens are not recognized by T-cells, they are degraded.

At the moment, we are able to predict the proteasomal cleavage of proteins into smaller fragments, and the binding of peptide sequences to MHC class I molecules. MAPPP first generates a probability for the cleavage of each possible peptide from a protein by the proteasome in the cell. This probability is based on a statistic-empirical method developed by H.-G. Holzhütter et al. Peptides with the highest probabilities are then given a score reflecting their abibility of binding to MHC molecules, which will transport the peptide to the cell surface. This binding score employs coefficient tables deduced from the literature by Kenneth Parker.
While there are some programs available for calculating a MHC-binding score, and a non-web-based sheet for the cleavage probability prediction was developed, none of them is able to handle a sequence of multiple proteins. We wanted to combine these two steps and have the possibility to process even an entire genome beginning with its translation into proteins, the cleavage prediction all the way through the MHC-binding scores.

Cleavage prediction

For calculating the cleavage probability of a specific fragment, we first determine the probability of a cut after each of the residues within the sequence. By limiting the minimum probability for a single residue in the form, one can limit the number of resulting possibilities. After that, a cleavage probability for all possible fragments between two cut-sites and with the right length (choosable in the form) is calculated. The cleavage probability for a fragment depends on the probability of either the N- and the C-residue, as well as the probabilities of the residues between these sites. The flanking regions to the left and the right of a fragment and the probabilities of their residues are considered too.

If you use the Proteasome cleavage prediction program by itself, some further data is shown, together with the output mentioned above. At the bottom, all the protein sequences are listed, together with the cleavage probability for each amino acid (as a hyperlink, just move the mouse on the residue; if the probability is below the specified minimum, none is shown) and the number of occurances in the peptide fragments for each amino acid. Beneath this information, you can find all the peptide fragments just below their corresponding amino acids in the sequence.

The program uses the algorithms first implemented in FRAGPREDICT, a computer program for the prediction of proteasomal cleavage sites and proteolytic fragments, developed by H.-G. Holzhütter et al.
Alternatively, you can choose that MAPPP uses the prediciton algorithms of PaProC, developed by H.-G. Rammensee et al.

MHC binding prediction

The prediction of the peptides binding to MHC class I molecules is based on a score calculated for each subsequence. Each specific amino acid at a specific position within a subsequence is given a value. Depending on the algorithms selected in the form, the values for the 8 to 10 amino acids are then multiplied (BIMAS) or added (SYFPEITHI) to determine the score for the subsequence.
These values have been pre-calculated and stored in static matrices. The pre-calculation was done by the BIMAS- and SYFPEITHI-staff.


J. Hakenberg, A. Nussbaum, H. Schild, H.-G. Rammensee, C. Kuttler, H.-G. Holzhütter, P.-M. Kloetzel, S.H.E. Kaufmann, H.-J. Mollenkopf (2003)
MAPPP - MHC-I Antigenic Peptide Processing Prediction. Applied Bioinformatics 2(3):155-158
H.-G. Holzhütter, C. Frömmel, and P.-M. Kloetzel (1999).
A Theoretical Approach Towards the Identification of Cleavage-Determining Amino Acid Motifs of the 20S Proteasome. J. Mol. Biol. 286, 1251-1265
H.-G. Holzhütter and P.-M. Kloetzel (2000).
A kinetic model of vertebrate 20S proteasome accounting for the generation of major proteolytic fragments from oligomeric peptide substrates. Biophysical J. 2000 (79): 1196-1205.
H.-G. Holzhütter et al.
Designed 1999 by the Institute of Biochemistry at the Medical School (Charité) of the Humboldt-University
C. Kuttler, A.K. Nussbaum, T.P. Dick, H.-G. Rammensee, H. Schild, K.P. Hadeler (2000).
An algorithm for the prediction of proteasomal cleavages, J. Mol. Biol. 298 (2000), 417-429
H.-G. Rammensee, J. Bachmann, N.N. Emmerich, O.A. Bachor, and S. Stevanovic (1999).
SYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics (1999) 50: 213-219
H.-G. Rammensee, J. Bachmann, S. Stevanovic (1997).
MHC ligands and peptide motifs. Landes Bioscience 1997 (International distributor - except North America: Springer Verlag GmbH & Co. KG, Tiergartenstr. 17, D-69121 Heidelberg)
H.-G. Rammensee, Friede, S. Stevanovic (1995).
MHC ligands and peptide motifs: 1st listing, Immunogenetics 41, 178-228
K. C. Parker, M. A. Bednarek, and J. E. Coligan (1994).
Scheme for ranking potential HLA-A2 binding peptides based on independent binding of individual peptide side-chains. J. Immunol. 152:163.
R. Taylor.
HLA Peptide Binding Predictions


Additionally supported by
World Health Organization
Fonds der chemischen Industrie
Hosted by
Association for Scientific Data Processing Göttingen


Stefan H.E. Kaufmann
Project leader
Hans Mollenkopf
Jörg Hakenberg
MHC1-Epitope Database
Sandra Lövenich
We would be very grateful for any feedback, comments and suggestions! Please send any correspondence to Jörg.
If you want to publish the results of the program, please cite:
Jörg Hakenberg, Alexander Nussbaum, Hansjörg Schild, Hans-Georg Rammensee, Christina Kuttler, Hermann-Georg Holzhütter, Peter-M. Kloetzel, Stefan H.E. Kaufmann, and Hans-Joachim Mollenkopf:
MAPPP - MHC-I Antigenic Peptide Processing Prediction.
Applied Bioinformatics, 2(3):155-158, 2003.
[Applied Bioinformatics] [PubMed]

Links to other interesting sites

Proteasome processing Peptide Binding Prediction DNA to protein translation

Further reading

ERAAP - The Aminopeptidase Associated with Antigen Processing in the Endoplasmatic Reticulum

Thomas Serwold, Federico Gonzalez, Jennifer Kim, Richard Jacob & Nilabh Shastri (2002).
ERAAP customizes peptides for MHC class I molecules in the endoplasmic reticulum. Nature 2002, Oct 3, 419(6906):480-483. DOI:10.1038/nature01074
Thomas Serwold, Stephanie Gaw & Nilabh Shastri (2001).
ER aminopeptidases generate a unique pool of peptides for MHC class I molecules. Nature Immunol, 2(7):644-651. DOI:10.1038/89800
Ian A. York, Shih-Chung Chang, Tomo Saric, Jennifer A. Keys, Janice M. Favreau, Alfred L. Goldberg & Kenneth L. Rock (2002).
The ER aminopeptidase ERAP1 enhances or limits antigen presentation by trimming epitopes to 8-9 residues. Nature Immunol, 3(12):1177-1184. DOI:10.1038/ni860
Tomo Saric, Shih-Chung Chang, Akira Hattori, Ian A. York, Shirley Markant, Kenneth L. Rock, Masafumi Tsujimoto & Alfred L. Goldberg (2002).
An IFN-γ-induced aminopeptidase in the ER, ERAP1, trims precursors to MHC class I-presented peptides. Nature Immunol, 3(12):1169-1176. DOI:10.1038/ni859