HLAIIPred: Cross-Attention Mechanism for Modeling the Interaction of HLA Class II Molecules with Peptides

The training, validation, and test data sets for the training of the HLAIIPred. The datasets are curated as described in the manuscript and from the following sources: MS-based experimental data from Racle et al.[1] and HLA Ligand Atlas[2], and monoallelic HLAII Immunopeptidomics dataset from Strazar et al.[3] Abstract --------- We introduce HLAIIPred, a deep learning model to predict peptides presented by class II human leukocyte antigens (HLAII) on the surface of antigen presenting cells. HLAIIPred is trained using a Transformer-based neural network and a dataset comprising of HLAII-presented peptides identified by mass spectrometry. In addition to predicting peptide presentation, the model can also provide important insights into peptide-HLAII interactions by identifying core peptide residues that form such interactions. We evaluate the performance of HLAIIPred on three different tasks, peptide presentation in monoallelic samples, immunogenicity prediction of therapeutic antibodies, and neoantigen prioritization for cancer immunotherapy. Additionally, we created a new dataset of biotherapeutics HLAII peptides presented by human dendritic cells. This data is used to develop screening strategies to predict the unwanted immunogenic segments of therapeutic antibodies by HLAII presentation models. HLAIIPred demonstrates superior or equivalent performance when compared to the latest models across all evaluated benchmark datasets. We achieve a 16% increase in prediction of presented peptides compared to the second-best model on a set of unseen peptides presented by less frequent alleles. The model improves clinical immunogenicity prediction, identifies epitopes in therapeutic antibodies and prioritize neoantigens with high accuracy. Zenodo record 15299217, published 2024-10-03, 3 file(s) totalling 242.8 MB (3 tabular).

Category
Proteomics
Provider
Mojtaba Haghighatlari
License
CC-BY-4.0
DOI
10.1101/2024.10.01.616078
Hosted on
Zenodo
Access
Free and public. No account with the data provider is needed.
Source
https://doi.org/10.1101/2024.10.01.616078
Data structure and loading code

Format: csv

Fields

peptidestring
presentedinteger
clusterinteger
weight_cdfloat
allele1string
allele2string
allele3string
allele4string
allele5string
allele6string
allele7string
allele8string
allele9string
allele10integer
allele11integer
allele12integer
allele13integer
allele14integer

Files

Example file: https://zenodo.org/api/records/15299217/files/train_positive.csv/content

Path pattern: Flat Zenodo record: every file hangs directly off the record, no directory tree.

Load the data

import pandas as pd

# train_positive.csv is 205.2 MB. Preview the first rows over HTTPS; Zenodo supports
# Range requests, so this reads a slice rather than the whole file.
url = "https://zenodo.org/api/records/15299217/files/train_positive.csv/content"
df = pd.read_csv(url, sep=",", nrows=5_000, low_memory=False)
print(df.shape, "(preview)")
# Full read, 205.2 MB into memory - uncomment when you mean it:
# df = pd.read_csv(url, sep=",", low_memory=False)
df.head()

Written for a Scigantic notebook, where /mnt/archive is the mounted dataset.

Analyze this dataset without downloading it

On Scigantic, HLAIIPred: Cross-Attention Mechanism for Modeling the Interaction of HLA Class II Molecules with Peptides opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

Related Proteomics datasets