Abstract:Epitope vaccines are a promising direction to enable precision treatment for cancer, autoimmune diseases, and allergies. Effectively designing such vaccines requires accurate prediction of proteasomal cleavage in order to ensure that the epitopes in the vaccine are presented to T cells by the major histocompatibility complex (MHC). While direct identification of proteasomal cleavage \emph{in vitro} is cumbersome and low throughput, it is possible to implicitly infer cleavage events from the termini of MHC-presented epitopes, which can be detected in large amounts thanks to recent advances in high-throughput MHC ligandomics. Inferring cleavage events in such a way provides an inherently noisy signal which can be tackled with new developments in the field of deep learning that supposedly make it possible to learn predictors from noisy labels. Inspired by such innovations, we sought to modernize proteasomal cleavage predictors by benchmarking a wide range of recent methods, including LSTMs, transformers, CNNs, and denoising methods, on a recently introduced cleavage dataset. We found that increasing model scale and complexity appeared to deliver limited performance gains, as several methods reached about 88.5% AUC on C-terminal and 79.5% AUC on N-terminal cleavage prediction. This suggests that the noise and/or complexity of proteasomal cleavage and the subsequent biological processes of the antigen processing pathway are the major limiting factors for predictive performance rather than the specific modeling approach used. While biological complexity can be tackled by more data and better models, noise and randomness inherently limit the maximum achievable predictive performance. All our datasets and experiments are available at <a class="link-external link-https" href="https://github.com/ziegler-ingo/cleavage_prediction" rel="external noopener nofollow">this https URL</a>.

Predicting Cleavage Sites in Exogenous Antigen Using Weighted SVM

Predicting Signal Peptides and Their Cleavage Sites Using Support Vector Machines and Improved Position Weight Matrixes

Prediction of T-cell epitopes based on least squares support vector machines and amino acid properties.

Proteasomal Cleavage Site Prediction of Protein Antigen Using BP Neural Network Based on a New Set of Amino Acid Descriptor

Predicting Immunogenic T-cell Epitopes by Combining Various Sequence-Derived Features.

Improved proteasomal cleavage prediction with positive-unlabeled learning

Studies on the Prediction of Selective Cleavage Sites and Cleavage Profile of Proteasome Using VHSE Amino Acid Descriptor

NIEluter: Predicting peptides eluted from HLA class I molecules

Prediction of MHC-binding Peptides of Flexible Lengths from Sequence-Derived Structural and Physicochemical Properties

A novel locally linear embedding and wavelet transform based encoding method for prediction of MHC-II binding affinity

A relevance vector machine based quantitative prediction method for mouse class i MHC peptide binding affinity

Quantitative prediction of MHC-II peptide binding affinity using relevance vector machine

What cleaves? Is proteasomal cleavage prediction reaching a ceiling?

Prediction of MHC class II Epitopes Using Fourier Analysis and Support Vector Machines

Quantitative Prediction of Mhc-Ii Peptide Binding Affinity Using Global Description of Peptide Sequences

Ranking-based Convolutional Neural Network Models for Peptide-MHC Binding Prediction

Prediction of MHC class i binding peptides using fourier analysis and support vector machine

Analysis of Specificity of Proteasome Cleavage Site Based on SVM Method

A New Kernel Based on Weighted Cross-Correlation Coefficient for SVMs and Its Application on Prediction of T-cell Epitopes

Prediction of t-cell epitopes using support vector machine and similarity kernel

Improving prediction of MHC class I Binding peptides using weighted SVM ensemble