… Two common approaches to determine prediction performance are cross-validation, in …
performances generated on our last benchmark dataset from 2009 with prediction performances …

… For each server, an overall ranking score is calculated, summarizing its overall performance
across all MHC molecules, peptide lengths and measurement data types. The ranking score …

… We firstly introduced the test set and evaluate the prediction performance of MHC class I
and II tools on the blind test set. The tools include published IEDB methods, MixMHCpred[17], …

… an automated platform to benchmark peptide-MHC class II binding prediction tools. The
platform evaluates the absolute and relative predictive performance of all participating tools on …

… performance benchmarking and assessment of currently available, state-of-the-art tools for
predicting peptide … RANKPEP [23] predicts the MHC class I-binding peptides using profile …

… When evaluating these three approaches, we found that the number of times the peptide
was identified by MS had the best performance with an AUC of 0.674 (AUC of combined score …

… of benchmark performance. We highly encourage anyone working on MHC binding predictions
to participate in this benchmark to … This work focuses on peptides bound to MHC class I …

… Specifically for the current dataset, we recommend evaluating prediction performance by
the ability to classify peptides into binders and nonbinders at a cutoff of 500 nM. We plan to …

… Weekly automated benchmarking is performed … performance evaluation of publicly available
MHC ligand prediction algorithms, based on a newly generated experimental MHC–peptide …

… the performance of several tools for side-chain prediction within the grooves of peptide-MHC …
The workflow of the benchmark is illustrated in Figure 4, and the single steps are described …