DeepPhage: leveraging the power of NGS phage display data, an in-depth analysis for the identification of Eu3+-binding peptides
- 1Helmholtz-Zentrum Dresden-Rossendorf Helmholtz Institute Freiberg for Resource Technology
Abstract
Phage Surface Display (PSD) is a widely used molecular technique for identifying peptides and antibodies with applications in therapeutics, diagnostics, and materials science. The use of highly diverse phage libraries enables the screening of millions of ligand variants, while the integration of Next-Generation Sequencing (NGS) provides high-throughput access to the sequence composition of selected phage populations. Despite these advances, the identification of genuine target-binding ligands remains challenging due to biases introduced during library construction, biological selection and propagation, as well as the enrichment of target-unrelated peptides (TUPs). These factors distort sequence distributions and complicate the interpretation of PSD-NGS experiments, particularly when comparing multiple libraries or selection rounds. Consequently, robust computational methods are required to distinguish biologically relevant enrichments from technical artifacts. We demonstrate DeepPhage using PSD-NGS datasets from Eu3+ biopanning experiments, in which differences in peptide composition between experimental conditions were identified and a systematic comparison of several conditions was made possible. It was developed to address the lack of comparative analysis tools for large-scale PSD-NGS experiments by combining statistical analyses with comprehensive visualisation. These features form the basis for a more reliable interpretation of phage display selection experiments. DeepPhage provides a reproducible framework for comparative PSD-NGS analysis, which supports the identification of sequence bias, library heterogeneity and potential target-binding peptides across multiple experiments. The software is freely available as an open-source command-line application, so that it can be deployed on a high computer cluster for particularly large volumes of data. Availability and implementation: DeepPhage is entirely open source and is available on GitLab https://codebase.helmholtz.cloud/phagelabs-hif/DeepPhage/-/tree/stable_release_v0.1 under the GPL v3 licence. The repository contains documentation, installation instructions, a user guide and example data. The sample data used for the demonstration is available at the online repository RODARE (https://doi.org/10.14278/rodare.4647).
Related
- Pair-complete structural gating enables Pareto optimization of de novo cyclic peptides with preserved binding geometry and improved developability
- Activation of Disease-Associated Polycystin-1 Mutants by Binding of Stalk-derived Agonist Peptides
- DeCTCF: Decoding CTCF binding sequences by leveraging predicted epigenomic features
- Identification and expression analysis of antibacterial peptides from venom of the scorpion Mesobuthus martensii
- Identification of RNA-Binding Proteins Regulating the Alternative Splicing of TREM2 Exon 4
BibTeX
@misc{bloss2026deepphage,
title = {{DeepPhage: leveraging the power of NGS phage display data, an in-depth analysis for the identification of Eu3+-binding peptides}},
author = {C. Bloss and A. Hübner and C. Renard and G. Techert and F. Lederer},
howpublished = {bioRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.29.762794},
url = {https://www.biorxiv.org/content/10.64898/2026.10.29.762794v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - DeepPhage: leveraging the power of NGS phage display data, an in-depth analysis for the identification of Eu3+-binding peptides
AU - C. Bloss
AU - A. Hübner
AU - C. Renard
AU - G. Techert
AU - F. Lederer
PY - 2026
DA - 2026/10/29
PB - bioRxiv
DO - 10.64898/2026.10.29.762794
UR - https://www.biorxiv.org/content/10.64898/2026.10.29.762794v1
AB - Phage Surface Display (PSD) is a widely used molecular technique for identifying peptides and antibodies with applications in therapeutics, diagnostics, and materials science. The use of highly diverse phage libraries enables the screening of millions of ligand variants, while the integration of Next-Generation Sequencing (NGS) provides high-throughput access to the sequence composition of selected phage populations. Despite these advances, the identification of genuine target-binding ligands remains challenging due to biases introduced during library construction, biological selection and propagation, as well as the enrichment of target-unrelated peptides (TUPs). These factors distort sequence distributions and complicate the interpretation of PSD-NGS experiments, particularly when comparing multiple libraries or selection rounds. Consequently, robust computational methods are required to distinguish biologically relevant enrichments from technical artifacts. We demonstrate DeepPhage using PSD-NGS datasets from Eu3+ biopanning experiments, in which differences in peptide composition between experimental conditions were identified and a systematic comparison of several conditions was made possible. It was developed to address the lack of comparative analysis tools for large-scale PSD-NGS experiments by combining statistical analyses with comprehensive visualisation. These features form the basis for a more reliable interpretation of phage display selection experiments. DeepPhage provides a reproducible framework for comparative PSD-NGS analysis, which supports the identification of sequence bias, library heterogeneity and potential target-binding peptides across multiple experiments. The software is freely available as an open-source command-line application, so that it can be deployed on a high computer cluster for particularly large volumes of data. Availability and implementation: DeepPhage is entirely open source and is available on GitLab https://codebase.helmholtz.cloud/phagelabs-hif/DeepPhage/-/tree/stable_release_v0.1 under the GPL v3 licence. The repository contains documentation, installation instructions, a user guide and example data. The sample data used for the demonstration is available at the online repository RODARE (https://doi.org/10.14278/rodare.4647).
ER -