iDeepLC: Chemical Structure Information Yields Improved Retention Time Prediction of Peptides with Unseen Modifications
Abstract
Deep learning has notably advanced the field of liquid chromatography-mass spectrometry-based proteomics. Accurate prediction of peptide retention times significantly enhances our ability to match LC-MS data with the correct peptides and proteins, especially for data-independent acquisition data. While numerous models predict peptide LC retention times with high accuracy, few can accurately predict the retention times of chemically modified peptides, particularly those with modifications not encountered during model training. In our previously developed DeepLC model, accurate predictions could be made for unseen modifications by leveraging the chemical compositions of (modified) residues. Here, however, we present a further enhancement of this model based on the chemical structural information. The resulting model, called iDeepLC, shows overall more accurate predictions and better generalization performance for predicting the retention time of modifications structurally defined as SMILES but unseen during training than DeepLC. iDeepLC is freely available as an open-source software under the Apache2 license and can be found at https://github.com/CompOmics/iDeepLC.
Related
- Reliable Structure-Processing-Performance Learning in Organic Photovoltaics: Chemical Generalization, Processing-Aware Prediction and Uncertainty under Domain Shift
- Bromo-desthiobiotin (Br-DTB) tag for unbiased profiling and identification of chemical modifications.
- New approach for deciphering polyethylene terephthalate (PET)-based plastics’ diols: Application on PET daily-life packaging objects paving their chemical modifications
- Ultrasound-mediated structural and chemical modifications in wheat proteins and their impact on functional and storage properties: a review.
- Solvent-Dependent Selective Retention of Essential-Oil Markers and Modification of the Thymol-Type Chemical Fingerprint in Post-Extraction Thymus vulgaris L. Biomass
BibTeX
@article{nameni2026ideeplc,
title = {{iDeepLC: Chemical Structure Information Yields Improved Retention Time Prediction of Peptides with Unseen Modifications}},
author = {Alireza Nameni and Arthur Declercq and Ralf Gabriels and Robbe Devreese and Sven Degroeve and Amélie de Maesschalck and Maarten Dhaenens and Cristina Chiva and Eduard Sabidó and Lennart Martens and Robbin Bouwmeester},
journal = {Analytical Chemistry},
year = {2026},
month = nov,
doi = {10.1021/acs.analchem.5c08017},
url = {https://hal.science/hal-05864770v1},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - iDeepLC: Chemical Structure Information Yields Improved Retention Time Prediction of Peptides with Unseen Modifications
AU - Alireza Nameni
AU - Arthur Declercq
AU - Ralf Gabriels
AU - Robbe Devreese
AU - Sven Degroeve
AU - Amélie de Maesschalck
AU - Maarten Dhaenens
AU - Cristina Chiva
AU - Eduard Sabidó
AU - Lennart Martens
AU - Robbin Bouwmeester
PY - 2026
DA - 2026/11/15
JO - Analytical Chemistry
DO - 10.1021/acs.analchem.5c08017
UR - https://hal.science/hal-05864770v1
AB - Deep learning has notably advanced the field of liquid chromatography-mass spectrometry-based proteomics. Accurate prediction of peptide retention times significantly enhances our ability to match LC-MS data with the correct peptides and proteins, especially for data-independent acquisition data. While numerous models predict peptide LC retention times with high accuracy, few can accurately predict the retention times of chemically modified peptides, particularly those with modifications not encountered during model training. In our previously developed DeepLC model, accurate predictions could be made for unseen modifications by leveraging the chemical compositions of (modified) residues. Here, however, we present a further enhancement of this model based on the chemical structural information. The resulting model, called iDeepLC, shows overall more accurate predictions and better generalization performance for predicting the retention time of modifications structurally defined as SMILES but unseen during training than DeepLC. iDeepLC is freely available as an open-source software under the Apache2 license and can be found at https://github.com/CompOmics/iDeepLC.
SN - 0003-2700
SN - 1520-6882
ER -