High-Throughput MHC I Ligand Prediction Using MHCflurry

Timothy O'Donnell; Alex Rubinsteyn

doi:10.1007/978-1-0716-0327-7_8

High-Throughput MHC I Ligand Prediction Using MHCflurry

Methods Mol Biol. 2020;2120:113-127. doi: 10.1007/978-1-0716-0327-7_8.

Authors

Timothy O'Donnell¹, Alex Rubinsteyn²

Affiliations

¹ Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, USA. tim@openvax.org.
² Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY, USA.

PMID: 32124315
DOI: 10.1007/978-1-0716-0327-7_8

Abstract

MHCflurry is an open source package for peptide/MHC I binding affinity prediction. Its command-line and programmatic interfaces make it well-suited for integration into high-throughput bioinformatic pipelines. Users can download models fit to publicly available data or train predictors on their own affinity measurements or mass spec datasets. This chapter gives a tutorial on essential MHCflurry functionality, including generating predictions, training new models, and using the MHCflurry Python interface. MHCflurry is available at https://github.com/openvax/mhcflurry .

Keywords: Epitope prediction; HLA; Immunoinformatics; MHC; Neoantigen.

MeSH terms

Computational Biology / methods
Histocompatibility Antigens Class I / metabolism*
Humans
Ligands
Peptides / chemistry
Peptides / metabolism*
Protein Binding
Software*

Substances

Histocompatibility Antigens Class I
Ligands
Peptides