iDeepLC: Chemical Structure Information Yields Improved Retention Time Prediction of Peptides with Unseen Modifications
Open Access
- 8 June 2026
- journal article
- Published by American Chemical Society (ACS) in Analytical Chemistry
- Vol. 98 (24) , 17707-17717
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.Keywords
Funding Information
- Research Foundation Flanders (FWO) (12A6L24N)
- Research Foundation Flanders (FWO) (12B7123N)
- Research Foundation Flanders (FWO) (1SE3724N)
- Research Foundation Flanders (FWO) (1SH9O24N)
- Research Foundation Flanders (FWO) (G010023N)
- Research Foundation Flanders (FWO) (G028821N)
- CERCA programme and The Departament de Recerca i Universitats de la Generalitat de Catalunya (2021SGR01225)
- European Union?s Horizon 2020 Programme (H2020-INFRAIA-2018-1)
- Spanish Infrastructure for Omics Technologies (ICTS OmicsTech)
- Spanish Ministry of Science, Innovation and Universities (CEX2020-001049-S)
- Spanish Ministry of Science, Innovation and Universities (PID2024-157336NB-I00)
- CHIST-ERA project ODEEP-EU (G0GDV23N)
- Horizon Europe Projects COMBINE (101191739)
- Horizon Europe Projects BAXERNA 2.0 (101080544)
- Ghent University Concerted Research Action (BOF21/GOA/033)
- European Union?s Horizon 2020 research and innovation programme (956148)
This publication has 45 references indexed in Scilit:
- A transformer architecture for retention time prediction in liquid chromatography mass spectrometry‐based proteomicsProteomics, 2023
- MS2Rescore: Data-Driven Rescoring Dramatically Boosts Immunopeptide Identification RatesMolecular & Cellular Proteomics, 2022
- INFERYS rescoring: Boosting peptide identifications and scoring confidence of database search resultsRapid Communications in Mass Spectrometry, 2021
- pDeep3: Toward More Accurate Spectrum Prediction with Fast Few-Shot LearningAnalytical Chemistry, 2021
- Assessing the Relationship Between Mass Window Width and Retention Time Scheduling on Protein Coverage for Data-Independent AcquisitionJournal of the American Society for Mass Spectrometry, 2019
- Peptide Retention Time Prediction in Hydrophilic Interaction Liquid Chromatography: Data Collection Methods and Features of Additive and Sequence-Specific ModelsAnalytical Chemistry, 2017
- Now, More Than Ever, Proteomics Needs Better ChromatographyCell Systems, 2016
- Ultradeep Human Phosphoproteome Reveals a Distinct Regulatory Nature of Tyr and Ser/Thr-Based SignalingCell Reports, 2014
- Probability-based protein identification by searching sequence databases using mass spectrometry dataElectrophoresis, 1999
- Neocognitron: A hierarchical neural network capable of visual pattern recognitionNeural Networks, 1988