MixMHC2pred is a predictor of MHC-II ligands displayed at the cell surface. It combines predictions of affinity to MHC-II molecules together with other features linked to antigen processing and presentation. It can work for any human MHC-II allele (HLA-DR/-DP/-DQ) as well as for many alleles from other species.
Peptides with lower %Ranks in MixMHC2pred predictions are more likely to elicit CD4+ T cell recognition
- Peptide and Context give the sequence of the peptides (respectively context), in the same order as provided in input.
- BestAllele gives the name of the best allele (based on the allele with best predicted score for the given peptide).
- %Rank_... give the predicted score for the best allele and for each allele asked. The score is given as a percentile rank, best score is about 0, worst score is 100. This value corresponds to the percent of random peptides that would have a better score than the peptide provided in input among peptides of sizes 12-21 amino acids.
- Core_best indicates the best predicted core binding sequence for each peptide towards its best allele.
- CoreP1_... give the most likely binding core position for the given peptide towards the given allele (this tells the position of the first amino acid from the binding core (which has a size of 9 aa in the predictions), starting at a value of 1 (i.e. if binding core corresponds to the 9 first amino acids from the peptide, this CoreP1 = 1)).
- subSpec_... tell in which sub-specificity the given peptide is likely bound toward the given allele. The value 1 corresponds to the main sub-specificity (the only one for multiple alleles). But for example for DRB1*08:01 allele a 2nd sub-specificity exists and is indicated by the value 2. For alleles accomodating reverse binding, a value of -1 indicates that the given peptide is bound in the reverse orientation.
Peptides that are too short (less than 12 amino acids), too long (more than 21 amino acids) or that contain non standard amino acids have NA values instead of their scores.
The best way to rank your peptides is to use the global score with the best allele (%Rank_best), the best predicted peptides have the lowest scores.
If you plan to use MixMHC2pred for commercial purposes, you are required to obtain a separate license. To do so, please contact Nadette Bulgin ([email protected]) at the Ludwig Institute for Cancer Research Ltd.
Julien Racle ([email protected]). Please provide as much details as possible and ideally send also an example input file that is causing the issue.
If you are using MixMHC2pred, please refer to
Racle, J., et al., Machine learning predictions of MHC-II specificities reveal alternative binding mode of class II epitopes. Immunity (2023) (available here).
and
Racle, J., et al. Robust prediction of HLA class II epitopes by deep motif deconvolution of immunopeptidomes. Nat. Biotechnol. 37, 1283–1286 (2019) (available here).