mhctools

mhctools is a Python library for running MHC binding, presentation, immunogenicity, and antigen-processing predictors. It provides a common interface to tools such as NetMHCpan and MHCflurry, with results you can inspect in Python or export as a pandas DataFrame.

Quickstart

Install mhctools and download the MHCflurry model weights:

pip install mhctools
mhctools fetch mhcflurry

from mhctools import MHCflurry

predictor = MHCflurry(alleles=["HLA-A*02:01"])
results = predictor.predict(["SIINFEKL", "GILGFVFTL"])
df = predictor.predict_dataframe(["SIINFEKL", "GILGFVFTL"])

The getting started guide explains the results and shows how to use another predictor.

Choose what to predict

Question	Predictors
Class I binding affinity	NetMHCpan or MHCflurry
Class I presentation	NetMHCpan, MHCflurry, MixMHCpred, BigMHC (EL), CapHLA
Class II binding or presentation	NetMHCIIpan; MixMHC2pred for presentation
Peptide-MHC complex stability	NetMHCstabpan
Proteasomal cleavage	Pepsickle or NetChop
Class II cleavage	NetCleave (class II)
TAP transport	DeepTAP
ERAP1 trimming	ERAMER
T-cell immunogenicity	Calis, PRIME, BigMHC (IM), DeepImmuno; TLimmuno2 for class II
Recognition by a specific TCR	NetTCR, Tulip, MixTCRpred
Descriptive foreignness	WEIRDO
Free-peptide half-life	Cavaco, PeptiVerse, PlifePred2
Cell penetration classification	PeptiVerse CPP
Per-bond peptidase evidence	Peptidase activity

The selection guide explains input, installation, and license constraints. Read each model's guide for its output and validation limits.

Start here

Use the recipes for protein scans, multiple samples and table annotation, or the API reference for class and function details. For installation help, see troubleshooting.

Read known limits and benchmarks when interpreting predictions. Research and evidence traces primary studies, validation audits and open calibration questions.

For processing workflows, start with the antigen-processing guide, then use batch assessments or vaccine reports.