Identification of MHC ligands through allele-guided isolation combined with machine learning for improved MHC assignment using ARDisplay-I
We’re pleased to share a new publication in Molecular & Cellular Proteomics (MCP) co-authored by Ardigen researchers, focused on improving MHC ligand identification and accurate assignment, an important step in T cell–based immunotherapies.
By combining optimized experimental approaches with machine learning, the study provides a more accurate way to interpret peptides displayed on the cell surfac
What’s the study about?
MHC ligand identification is essential for discovering targets for immunotherapies, but it remains challenging – particularly when peptides are found at low abundance on the cell surface.
In this study, the team:
- showed how isolation strategy and separation conditions directly influence which MHC ligands are recovered, down to individual epitopes
- demonstrated that peptide detection depends on properties such as hydrophobicity, post-translational modifications, and the presenting HLA allele
- developed ARDisplay-I, a machine learning model that improves assignment of peptides to specific MHC alleles
ARDisplay-I was benchmarked against established tools such as netMHCpan, MixMHCpred, and MHCflurr and outperformed them in predicting MHC class I ligand presentation.
Study significance
Selecting the right targets is one of the most critical steps in T cell–based immunotherapies – and it depends directly on how well MHC-bound peptides can be captured and interpreted.
Both experimental design and computational modeling play a role here. Combining optimized isolation strategies with machine learning improves both the visibility of relevant peptides and confidence in identifying which MHC alleles present them, especially in complex, multiallelic samples.
Together, these advances support a more reliable transition from experimental data to actionable therapeutic targets.
Authors
We would like to congratulate all co-authors on this achievement and thank everyone involved for their valuable contributions:
Shima Mecklenbräuker, Piotr Skoczylas, Paweł Biernat, Badeel K.H.Q. Zaghla, Bilge Atay, Mai Hossam, Bartłomiej Król-Józaga, Maciej Jasiński, Victor Murcia Pienkowski, Anna Sanecka-Duin, Oliver Popp, Mohamed Haji, Rafał Szatanek, Philipp Mertins, Jan Kaczmarczyk, Ulrich Keller, Agnieszka Blum, Martin G. Klatt.
