Biology subjects

Renard, B. Y.

Publications and source records attributed to Renard, B. Y..

At least 19 recordsLinked to original sources

bioRxiv2026

A High-Confidence Atlas of Protein Methylation Enables AI-Driven Detection of Methylated Peptides

Lysine and arginine methylation regulate chromatin dynamics, transcription, and cellular signaling, however confident mass spectrometry (MS)-based detection and localization of this modification remain challenging. We reanalyzed eight public human methylation-enriched datasets using an open and standardized workflow that integrates database searching via the Trans-Proteomic Pipeline with a decoy-based statistical method for the independent estimation of false localization rates (FLR). This yielded a high-confidence Human Methylation Atlas of 1,828 sites (57 methyl-lysine, 1,771 methyl-arginine) across 1,021 proteins, classified into Gold, Silver, and Bronze confidence tiers. This is far fewer sites than reported in previous studies, reflecting the application of stringent FLR control, and what we hypothesise is potential high-false discovery in previous analyses. We then leveraged this resource to adapt a deep learning-based methodology for the improved detection of methylated peptides. Three mouse methylation-enriched datasets were reanalysed to augment training, and the phosphoproteomics-trained AHLF (ad hoc learning of peptide fragmentation) model was fine-tuned by transfer learning to create AHLF-Methylation. The model achieved mean ROC-AUC values of 0.824 on human spectra, and 0.829 on combined human-mouse spectra. The atlas is available through PTMeXchange and PRIDE, with curated site evidence integrated into UniProt and PeptideAtlas.

bioinformatics↗

bioRxiv2026

ProteoDUDes: Taxonomic profiling for metaproteomics with false positive reduction

Metaproteomics is the investigation of the protein composition of multi-organism samples. While metagenomics answers the question which organisms are present in a sample, metaproteomics additionally answers the question which organisms are active. State-of-the-art tools for annotating proteomic data with taxonomic information (e.g. Unipept, DIAMOND) do not control the false taxonomic identification rate, which can lead to incorrect results and thus incorrect interpretations, as we demonstrate with examples. ProteoDUDes processes the results from popular sequence annotation tools so that the proportion of true identifications in the result is at as high as or higher than in the the compared tools. We evaluate ProteoDUDes on simulated data and experimental mock community data. Our results indicate that ProteoDUDes has the same error rate as other tools on simulated data and half the error rate on the experimental mock community data. This allows more accurate statements to be made about which organisms are functionally active in a complex sample. ProteoDUDes is open-source and available at https://github.com/pirovc/dudes.

bioinformatics↗

bioRxiv2026

A High-Quality Acetylation Dataset Reveals Modest Data Requirements for Transfer Learning to Identify Little Studied Post-Translational Modifications

Dysregulation of post-translational modifications (PTMs) is associated with severe pathologies, including cancers and Alzheimers disease. Despite their biological importance, identifying modified peptides remains challenging due to the immense combinatorial search space. While searches benefit from prior knowledge of a peptides modification status, the data scarcity for most PTMs hinders the development of accurate deep learning classifiers like AHLF (ad hoc learning of peptide fragmentation). Here, we overcome this data bottle-neck for acetylation and ubiquitination. We harmonised a dataset with about 500,000 high quality acetylated peptide-spectrum matches (PSMs) from nine publicly available acetylation-enriched datasets. We fine-tuned AHLF with the acetylation and a 2-million spectra strong ubiquitination dataset separately and assessed the minimum data requirement for training by iteratively downsampling. Training separate models on SILAC and label-free subsets also assessed the impact of data diversity. The resulting acetylation and ubiquitination models achieve an AUC of 0.87 and 0.90 respectively. Beyond 28,500 acetylated spectra, corresponding to roughly 0.3% of the original models training data, additional data just provides minor performance gains. Finally, we show that data diversity is beneficial for generalizability, while models trained on homogeneous data sources tend to overfit to their respective data type. All code, and model weights are available at https://gitlab.com/dacs-hpi/ahlf-ptmai.

bioinformatics↗

bioRxiv2026

A framework for peptide identification on commercial nanopore sequencing platforms

Direct single-molecule peptide analysis could in principle enable rapid and sensitive identification of pathogen-derived or disease-associated biomarkers without reliance on mass spectrometry. However, existing nanopore peptide sensing methods are typically constrained by limited throughput and lack of accessibility beyond specialized setups. Here, we present an integrated experimental-computational framework for DNA-linked peptide translocation on a commercially available, high-throughput nanopore sequencing platform, the MinION. Synthetic peptides were covalently bound to oligonucleotides at both termini. The resulting peptide-DNA constructs were then translocated through the CsgG-CsgF pores using a DNA motor protein. Current traces were segmented using the known DNA sequences to extract peptide-associated signal regions. From these segments, we extracted signal features and trained feature-based and deep-learning classifiers to distinguish peptides, balancing interpretability and classification performance. We establish a framework for peptide identification using standard nanopore sequencing hardware. Across a diverse panel of synthetic peptides, our approach resolves single-amino-acid substitutions, maintains performance across independent sequencing runs, and correctly identifies peptides in blind mixtures. Interpretable model analyses connect classifier decisions and common errors to specific signal motifs. By combining commercially available instrumentation with a reproducible experimental and computational workflow, this framework lowers the barrier to nanopore-based proteomics and enables broader adoption across laboratories. It provides a foundation for future developments in amino acid modification detection and sequence analysis.

bioinformatics↗

bioRxiv2026

Leveraging Uncertainty Estimates for Drug Response Prediction in Cancer Cell Lines

Machine learning models for drug response prediction in cancer cell lines carry the potential to advance precision oncology by tailoring treatments to the molecular tumor profile. Their application is challenged by variability in prediction quality and distribution shifts between training and application. Uncertainty estimation provides more information on the predictive distribution than point estimates, enabling comprehensive decision support and downstream analysis of predictions. Yet, the most effective uncertainty estimator is domain-specific. In this work, we benchmark uncertainty-aware models for drug response prediction. We focus on epistemic uncertainty via ensemble agreement, and aleatoric uncertainty via distributional modeling, or both. We find that ensemble-based estimates are more sensitive to distribution shift and can flag out-of-distribution examples. In contrast, distributional models yield stronger prediction error reductions among high-confidence subsets. Despite higher computational cost, the combination can provide both advantages: an ensemble of neural networks that estimate a Gaussian predictive distribution can reduce the mean squared error by 64 percent when restricting predictions to the 10 percent most confident drug-cell line pairs, and reliably indicates distribution shifts and platform differences. Beyond benchmarking, probabilistic predictions can identify drugs whose uncertainty bounds overlap with therapeutically relevant ranges. We also show that uncertainty estimates enable a new dimension of model interpretability: by attributing predicted uncertainty to input features, we identify genes that signal unpredictability of drug response rather than sensitivity or resistance. We further demonstrate uncertainty-guided selection of measurements for active learning. In summary, including uncertainty in drug response prediction supports better-informed model application. The code is available at https://github.com/PascalIversen/LUDRP.

bioinformatics↗

bioRxiv2026

usiGrabber: Automating the curation of proteomics spectra data at scale, making large datasets ready for use in machine learning systems

MotivationAn unprecedented amount of mass spectrometry-based proteomics data is publicly available through repositories such as the PRoteomics IDEntifications Database (PRIDE), and the field is increasingly leveraging machine learning approaches. However, the available data is not ready to be reused in a scalable way beyond the original acquisition purpose. Existing machine learning models commonly rely on a few manually curated datasets that require deep domain expertise and tedious technical work to construct. Importantly, these datasets have not been updated in recent years, so that newly published data remains inaccessible. We present usiGrabber, a scalable framework for assembling large proteomic datasets. usiGrabber is designed around portability and extensibility. It extracts spectra identification data from mzIdentML files, stores additional project-level metadata retrieved through the PRIDE API, indexes raw spectra using Universal Spectrum Identifiers (USIs), and offers download utilities to retrieve spectra data at scale. ResultsWithin 49 hours, we parsed over 800 million peptide spectrum matches and corresponding USIs from over 1,200 projects. As a proof of concept, we used usiGrabber to construct a phosphorylation-specific training dataset of nearly 11 million spectra in under two days and used it to retrain a binary phosphorylation classifier based on the AHLF model architecture. With a balanced accuracy of 0.78, our model achieves comparable performance to the original model on an independent test set, showing that automated data extraction is an alternative to manual curation of static datasets. AvailabilityAll code is available at https://github.com/usiGrabber/usiGrabber; the data is available at https://zenodo.org/records/18853258.

bioinformatics↗

bioRxiv2025

CombDNF: Disease-specific drug combination predictions with network-based features on clinically validated data

Drug combinations are increasingly applied to treat a wide range of complex diseases. Drug action and thus also drug combination effects can differ between diseases, e.g., due to molecular differences. Therefore, disease-specific predictions are required for treatments. A plethora of methods based on cell-line screening data in cancer has been proposed. However, their extendability to other diseases is limited, as is their applicability in the clinical context due to the in-vivo-in-vitro gap. In contrast, only few approaches rely on clinically validated data. Here, we propose CombDNF, a novel machine-learning based method for disease-specific drug combination prediction on clinically validated data. CombDNF is trained and predicts both on clinically approved (effective) and clinically reported adverse drug combinations from a broad collection of data sources. It can cope with the highly imbalanced label distribution in drug combination data. Further, CombDNF leverages network-derived features based on drug target and disease gene relationships. To incorporate uncertainty of the network topology it relies on edge weights in the underlying network. We systematically evaluate CombDNF against available state-of-the-art methods in four diseases with different underlying mechanisms and ground truth data characteristics. CombDNF outperforms all state-of-the-art methods in all four diseases by at least 84% in the AUPR. This translates, for example, to an enrichment of effective drug combinations in the top ten hypertension-specific predictions of four, compared to one for the best competing method. In addition, network edge weighting by interaction confidence scores indeed yields improved predictions. Further, we find evidence for biological plausibility of our top ranked drug combinations. We make the training and evaluation pipeline for CombDNF available ready-to-use at https://github.com/DILiS-lab/CombDNF.

bioinformatics↗

bioRxiv2024

End-to-end simulation of nanopore sequencing signals with feed-forward transformers

MotivationNanopore sequencing represents a significant advancement in genomics, enabling direct long-read DNA sequencing at the single-molecule level. Accurate simulation of nanopore sequencing signals from nucleotide sequences is crucial for method development and for complementing experimental data. Most existing approaches rely on predefined statistical models, which may not adequately capture the properties of experimental signal data. Furthermore, these simulators were developed for earlier versions of nanopore chemistry, which limits their applicability and adaptability to the latest flow cell data. ResultsTo enhance the quality of artificial signals, we introduce seq2squiggle, a novel transformer-based, non-autoregressive model designed to generate nanopore sequencing signals from nucleotide sequences. Unlike existing simulators that rely on static k-mer models, our approach learns sequential contextual information from the signal data. We benchmark seq2squiggle against state-of-the-art simulators on real experimental R10.4.1 data, evaluating signal similarity, basecalling accuracy, and variant detection rates. Seq2squiggle consistently outperforms existing tools across multiple datasets, demonstrating superior similarity to real data and offering a robust solution for simulating nanopore sequencing signals with the latest flow cell generation. Availability and Implementationseq2squiggle is freely available on GitHub at: github.com/ZKI-PH-ImageAnalysis/seq2squiggle

bioinformatics↗

bioRxiv2024

ema-tool: a Python Library for the Comparative Analysis of Embeddings from Biomedical Foundation Models

Foundation models, which encode patterns in large, high-dimensional data as embeddings, show promise in many machine learning related applications in molecular biology. Embeddings learned by the models provide informative features for downstream prediction tasks, however, the information captured by the model is often not interpretable. One approach to understanding the captured information is through the analysis of their learned embeddings, which in molecular biology so far has mainly focused on visualizing individual embedding spaces. This study introduces a quantitative framework for cross-space comparison, enabling intuitive exploration and comparison of embedding spaces in molecular biology. The framework emphasizes analyzing the distribution of known biological information within embedding space neighborhoods and provides insights into relationships between multiple embedding spaces. Comparison techniques include global pairwise distance measurements as well as local nearest neighbor analyses. By applying our framework to embeddings from protein language models, we demonstrate how embedding space analysis can serve as a valuable pre-filtering step for task-specific supervised machine learning applications and for the recognition of differential patterns in data encoded within and across different embedding spaces. To support a wide usability, we provide a Python library that implements all analysis methods, available at https://github.com/broadinstitute/EmmaEmb.

bioinformatics↗

bioRxiv2024

Optimizing ODE-derived Synthetic Data for Transfer Learning in Dynamical Biological Systems

BackgroundAccurate prediction of the temporal dynamics of biological systems is crucial for informing timely and effective interventions, e.g., in ecological or epidemiological contexts, or for treatment adjustments in therapy. While machine learning has proven its capabilities in generalizing the underlying non-linear dynamics of such systems, unlocking its predictive power is often restrained by the limited availability of large, curated datasets. To supplement real-world data, informing machine learning by transfer learning with synthetic data derived from simulations using ordinary differential equations has emerged as a promising solution. However, the success of this approach highly depends on the designed characteristics of the synthetic data. ResultsWe suggest scrutinizing these characteristics, such as size, diversity, and noise, of ordinary differential equation-based synthetic time series datasets. Here, we demonstrate how to systematically evaluate the influence of such design choices on transfer learning performance. We conduct a proof-of-concept study on three simple, but widely used systems and four real-world datasets. We find a strong interdependency between synthetic dataset size and diversity effects. Good transfer learning settings heavily rely on real-world data characteristics as well as the datas coherence with the dynamics of the model underlying the synthetic data. We achieve a performance improvement of up to 92% in mean absolute error for simulation-based transfer learning compared to non-informed deep learning. ConclusionsOur work emphasizes the relevance of carefully selecting properties of synthetic data for leveraging the valuable domain knowledge contained in ordinary differential equation models for machine-learning based predictions. The code is available at https://github.com/DILiS-lab/opt-synthdata-4tl.

bioinformatics↗

bioRxiv2024

PEPerMINT: Peptide Abundance Imputation in Mass Spectrometry-based Proteomics using Graph Neural Networks

MotivationAccurate quantitative information about the protein abundance is crucial for understanding a biological system and its dynamics. Protein abundance is commonly estimated using label-free, bottom-up mass spectrometry protocols. Here, proteins are digested into peptides before quantification via mass spectrometry. However, missing peptide abundance values, which can make up more than 50% of all abundance values, are a common issue. They result in missing protein abundance values, which then hinder accurate and reliable downstream analyses. ResultsTo impute missing abundance values, we propose PEPerMINT, a graph neural network model working directly on the peptide level that flexibly takes both peptide-to-protein relationships in a graph format as well as amino acid sequence information into account. We benchmark our method against eleven common imputation methods on six diverse datasets, including cell lines, tissue, and plasma samples. We observe that PEPerMINT consistently outperforms other imputation methods. Its prediction performance remains high for varying degrees of missingness, different evaluation approaches and differential expression prediction. As an additional novel feature, PEPerMINT provides meaningful uncertainty estimates and allows for tailoring imputation to the users needs based on the reliability of imputed values. Availability and implementationThe code is available at https://github.com/DILiS-lab/pepermint.

bioinformatics↗

bioRxiv2024

Motif Interactions Affect Post-Hoc Interpretability of Genomic Convolutional Neural Networks

Post-hoc interpretability methods are commonly used to understand decisions of genomic deep learning models and reveal new biological insights. However, interactions between sequence regions (e.g. regulatory elements) impact the learning process as well as interpretability methods that are sensitive to dependencies between features. Since deep learning models learn correlations between the data and output that do not necessarily represent a causal relationship, it is difficult to say how well interacting motif sets are fully captured. Here, we investigate how genomic motif interactions influence model learning and interpretability methods by formalizing possible scenarios where interaction effects appear. This includes the choice of negative data and non-additive effects on the outcome. We generate synthetic data containing interactions for those scenarios and evaluate how they affect the performance of motif detection. We show that post-hoc interpretability methods can miss motifs if interactions are present depending on how negative data is defined. Furthermore, we observe differences in interpretability between additive and non-additive effects as well as between post-hoc interpretability methods.

bioinformatics↗

bioRxiv2023

Taxor: Fast and space-efficient taxonomic classification of long reads with hierarchicalinterleaved XOR filters

Metagenomic long-read sequencing is gaining popularity for various applications, including pathogen detection and microbiome studies. To analyze the large data created in those studies, software tools need to taxonomically classify the sequenced molecules and estimate the relative abundances of organisms in the sequenced sample. Due to the exponential growth of reference genome databases, the current taxonomic classification methods have large computational requirements. This issue motivated us to develop a new data structure for fast and memoryefficient querying of long reads. Here we present Taxor as a new tool for long-read metagenomic classification using a hierarchical interleaved XOR filter data structure for indexing and querying large reference genome sets. Taxor implements several k-mer-based approaches such as syncmers for pseudoalignment to classify reads and an Expectation-Maximization algorithm for metagenomic profiling. Our results show that Taxor outperforms competing shortand long-read tools regarding precision, while having a similar recall. Most notably, Taxor reduces the memory requirements and index size by more than 50% and is among the fastest tools regarding query times. This enables real-time metagenomics analysis with large reference databases on a small laptop in the field. Taxor is available at https://gitlab.com/dacs-hpi/taxor.

bioinformatics↗

bioRxiv2023

Predicting Decision-Making Time for Diagnosis over NGS Cycles: An Interpretable Machine Learning Approach

MotivationGenome sequencing processes are commonly followed by computational analysis in medical diagnosis. The analyses are generally performed once the sequencing process has finished. However, in time-critical applications, it is crucial to start diagnosis once sufficient evidence has been accumulated. This research aims to define a proof-of-principle for predicting earlier time for decision-making using a machine learning approach. The method is evaluated on Illumina sequencing cycles for pathogen diagnosis. ResultsWe utilized a Long-Short Term Memory (LSTM) approach to make predictions for the early decision-making time in time-critical clinical applications. We modeled the (meta-)information obtained from NGS intermediate cycles to investigate whether there are any changes to expect in the remaining sequencing cycles. We tested our model on different patient datasets, resulting in high accuracy of over 98%, indicating the model is independent of a dataset. Furthermore, we can save several hours of turnaround time by using the early prediction results. We used the SHapley Additive exPlanations (SHAP) framework for the interpretation and assessment of the LSTM classifier. AvailabilityThe source code is available at https://gitlab.com/dacs-hpi/ngs-biclass. ContactBernhard.Renard@hpi.de

bioinformatics↗

bioRxiv2023

Evaluating Prompt Tuning for Conditional Protein Sequence Generation

Designing artificial proteins with specialized functions promises new solutions for biological, medical, and environmental use cases. This field benefits from advances in natural language processing, with state-of-the-art text generation models already being successfully applied to protein sequences. Openly available pre-trained protein language models are able to generate artificial protein sequences and can be finetuned on very specific tasks. Considering the high computational cost of finetuning a model exclusively for one downstream task, prompt tuning has been proposed as a more cost-efficient alternative that shares one model across different tasks. However, no openly available implementation of this approach compatible with protein language models has been previously published. Thus, we adapt an open-source codebase designed for NLP models to build a pipeline for prompt tuning on protein sequence data, supporting the protein language models ProtGPT2 and RITA. We benchmark this implementation for generating proteins of a specific family and evaluate the approach using text processing metrics as well as family membership prediction and protein activity prediction of generated sequences. Our results confirm the advantages of prompt tuning in resource usage, especially storage, encouraging further research and expansion of this technique to related use cases. For our evaluated use case, prompt tuning does not reach up to finetuning in terms of the quality of generated protein sequences, indicating the need for more extensive optimization. Lastly, we observe discrepancies between results of similar evaluation tools, highlighting open problems for principled assessment of protein sequence generation quality.

bioinformatics↗

bioRxiv2023

MpoxRadar: a worldwide Mpox genomic surveillance dashboard

Monkeypox (Mpox) is mutating at an exceptional rate for a DNA virus and its global spread is concerning, making genomic surveillance a necessity. With MpoxRadar, we provide an interactive dashboard to track virus variants on mutation level worldwide. MpoxRadar allows users to select among different genomes as reference for comparison. The occurrence of mutation profiles based on the selected reference is indicated on an interactive world map that shows the respective geographic sampling site in customizable time ranges to easily follow the frequency or trend of defined mutations. Furthermore, the user can filter for specific mutations, genes, countries, genome types, and sequencing protocols and download the filtered data directly from MpoxRadar. On the server, we automatically download all Mpox genomes and metadata from the National Center for Biotechnology Information (NCBI) on a daily basis, align them with the different reference genomes, generate mutation profiles, which are stored and linked to the available metainformation in a database. This makes MpoxRadar a practical tool for the genomic survaillance of Mpox, supporting users with limited computational resources. MpoxRadar is open-source and freely accessible at https://MpoxRadar.net. GRAPHICAL ABSTRACT O_FIG_DISPLAY_L [Figure 1] M_FIG_DISPLAY C_FIG_DISPLAY

bioinformatics↗

bioRxiv2022

In-silico enrichment of bacterial plasmids by chromosomal depletion using nanopore adaptive sampling

Bacterial plasmids play a major role in the spread of antibiotic resistance genes. However, their characterization via DNA sequencing suffers from the low abundance of plasmid DNA in those samples. Although sample preparation methods can enrich the proportion of plasmid DNA before sequencing, these methods are expensive and laborious, and they might introduce a bias by enriching only for specific plasmid DNA sequences. Nanopore adaptive sampling could overcome these issues by rejecting uninteresting DNA molecules during the sequencing process. In this study, we assess the application of adaptive sampling for the enrichment of low-abundant plasmids in known bacterial isolates using two different adaptive sampling tools. We show that a significant enrichment can be achieved even on expired flow cells. By applying adaptive sampling, we also improve the quality of de novo plasmid assemblies and reduce the sequencing time. However, our experiments also highlight issues with adaptive sampling if target and non-target sequences span similar regions.

genomics↗

bioRxiv2022

Current state, existing challenges, and promising progress for de novo sequencing and assembly of monoclonal antibodies

Monoclonal antibodies (mAbs) are biotechnologically produced proteins with various applications in research, therapeutics, and diagnostics. Their ability to recognize and bind to specific molecule structures makes them essential research tools and therapeutic agents. Sequence information of antibodies is helpful for understanding antibody-antigen interactions and ensuring their affinity and specificity. De novo protein sequencing based on mass spectrometry is a useful method to obtain the amino acid sequence of peptides and proteins without a priori knowledge. Deep learning-based approaches have been developed and applied more frequently to increase the accuracy of de novo sequencing. In this study, we evaluated five recently developed de novo sequencing algorithms (Novor, pNovo 3, DeepNovo, SMSNet, and PointNovo) in their ability to identify and assemble antibody sequences. The deep learning-based tools PointNovo and SMSNet showed an increased peptide recall across different enzymes and datasets compared to spectrum-graph-based approaches. We evaluated different error types of de novo peptide sequencing tools and their performance for different numbers of missing cleavage sites, noisy spectra, and peptides of various lengths. We achieved a sequence coverage of 93.15% to 99.07% on the light chains of three different antibody datasets using the de Bruijn assembler ALPS and the predictions from PointNovo. However, low sequence coverage and accuracy on the heavy chains demonstrate that complete de novo protein sequencing remains a challenging issue in proteomics that requires improved de novo error correction, alternative digestion strategies, and hybrid approaches such as homology search to achieve high accuracy on long protein sequences.

bioinformatics↗