Biology subjects

De Waele, G.

Publications and source records attributed to De Waele, G..

6 recordsLinked to original sources

bioRxiv2026

Reliable Molecular Retrieval from Mass Spectra using Conformal Prediction

A key task in the computational analysis of liquid chromatography-tandem mass spectrometry (LC- MS/MS) data is identifying the molecular structure underlying a measured spectrum. A common approach ranks candidate molecules retrieved from chemical databases using predicted fingerprint similarities, yet standard metrics such as top-k accuracy summarize performance only at the dataset level and provide no spectrum-specific reliability statement. In this work, we apply conformal prediction to candidate-based molecular retrieval to construct spectrum-specific prediction sets that contain the true molecule with a user-specified probability. We evaluate marginal and conditional conformal prediction across three experimental scenarios representing in-distribution, partially shifted, and fully out-of-distribution settings on the MassSpecGym benchmark. When calibration and test data are aligned, conformal prediction attains the target coverage with small candidate sets for most spectra. Under distribution shift, prediction sets become larger as rankings grow more ambiguous, although candidates can still be reduced when calibration remains representative. Conditional conformal prediction improves subgroup reliability across spectra of different difficulty, with the best gains obtained using confidence-based grouping. Overall, conformal prediction turns candidate rankings into reliable, spectrum-specific candidate sets with an explicit reliability-efficiency trade-off.

bioinformatics↗

bioRxiv2025

Harnessing machine learning to predict antibiotic susceptibility in Pseudomonas aeruginosa biofilms

Standard antibiotic susceptibility tests (ASTs) often fail to accurately predict treatment outcomes because they do not account for biofilm-specific mechanisms of reduced susceptibility. In the present study, we explored alternative approaches to predict tobramycin susceptibility of Pseudomonas aeruginosa biofilms that were experimentally evolved in physiologically relevant conditions. To this end, we used four analytical methods - whole-genome sequencing (WGS), matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS), isothermal microcalorimetry (IMC) and multi-excitation Raman spectroscopy (MX-Raman). Machine learning models were trained on data outputs from these methods to predict tobramycin susceptibility of our evolved strains and subsequently validated with a collection of clinical isolates. For minimal inhibitory concentration (MIC) predictions of the evolved strains, the highest accuracy{+/-}1 was achieved with MALDI-TOF MS (97.83 %), while for biofilm prevention concentration (BPC) predictions, Raman spectroscopy performed best with an accuracy{+/-}1 of 80.43 %. Overall, all analytical methods demonstrated comparable predictive performance, showing their potential for improving biofilm AST.

microbiology↗

bioRxiv2025

A systematic assessment of single-cell language model configurations

Transformers pre-trained on single-cell transcriptomic data have recently been applied to a series of tasks, earning them the title of foundation models. As all currently published models in this class employ vastly different pre-training strategies, it is impossible to determine which practices drive their success (or failure). Here, we present a large-scale study of pre-training components for single-cell transcriptomic transformers: bento-sc (BENchmarking Transformer-Obtained Single-Cell representations). By isolating (and tuning) parts of the pre-training scheme one by one, we define best practices for single-cell language model (scLM) construction. While comparisons with baselines indicate that scLMs do not yet offer the generational leap in prediction performances promised by many foundation models, we identify key design choices leading to their improved performance. Namely, the best scLMs are obtained by: (1) minimally processing counts at the input level, (2) using reconstruction losses that exploit known count distributions, (3) masking (up to high rates), and (4) combining different pre-training tasks/losses. All code supporting this study is distributed on PyPI and is packaged under: https://github.com/gdewael/bento-sc.

bioinformatics↗

bioRxiv2024

Pre-trained Maldi Transformers improve MALDI-TOF MS-based prediction

For the last decade, matrix-assisted laser desportion/ionization time-of-flight mass spectrometry (MALDI-TOF MS) has been the reference method for species identification in clinical microbiology. Hampered by a historical lack of open data, machine learning research towards models specifically adapted to MALDI-TOF MS remains in its infancy. Given the growing complexity of available datasets (such as large-scale antimicrobial resistance prediction), a need for models that (1) are specifically designed for MALDI-TOF MS data, and (2) have high representational capacity, presents itself. Here, we introduce Maldi Transformer, an adaptation of the state-of-the-art transformer architecture to the MALDI-TOF mass spectral domain. We propose the first self-supervised pre-training technique specifically designed for mass spectra. The technique is based on shuffling peaks across spectra, and pre-training the transformer as a peak discriminator. Extensive benchmarks confirm the efficacy of this novel design. The final result is a model exhibiting state-of-the-art (or competitive) performance on downstream prediction tasks. In addition, we show that Maldi Transformers identification of noisy spectra may be leveraged towards higher predictive performance. All code supporting this study is distributed on PyPI and is packaged under: https://github.com/gdewael/maldi-nn

bioinformatics↗

bioRxiv2023

An antimicrobial drug recommender system using MALDI-TOF MS and dual-branch neural networks

Timely and effective use of antimicrobial drugs can improve patient outcomes, as well as help safeguard against resistance development. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is currently routinely used in clinical diagnostics for rapid species identification. Mining additional data from said spectra in the form of antimicrobial resistance (AMR) profiles is, therefore, highly promising. Such AMR profiles could serve as a drop-in solution for drastically improving treatment efficiency, effectiveness, and costs. This study endeavours to develop the first machine learning models capable of predicting AMR profiles for the whole repertoire of species and drugs encountered in clinical microbiology. The resulting models can be interpreted as drug recommender systems for infectious diseases. We find that our dual-branch method delivers considerably higher performance compared to previous approaches. In addition, experiments show that the models can be efficiently fine-tuned to data from other clinical laboratories. MALDI-TOF-based AMR recommender systems can, hence, greatly extend the value of MALDI-TOF MS for clinical diagnostics. All code supporting this study is distributed on PyPI and is packaged under: https://github.com/gdewael/maldi-nn

bioinformatics↗

bioRxiv2021

CpG Transformer for imputation of single-cell methylomes

AO_SCPLOWBSTRACTC_SCPLOWO_ST_ABSMotivationC_ST_ABSThe adoption of current single-cell DNA methylation sequencing protocols is hindered by incomplete coverage, outlining the need for effective imputation techniques. The task of imputing single-cell (methylation) data requires models to build an understanding of underlying biological processes. ResultsWe adapt the transformer neural network architecture to operate on methylation matrices through combining axial attention with sliding window self-attention. The obtained CpG Transformer displays state-of-the-art performances on a wide range of scBS-seq and scRRBS-seq datasets. Further-more, we demonstrate the interpretability of CpG Transformer and illustrate its rapid transfer learning properties, allowing practitioners to train models on new datasets with a limited computational and time budget. Availability and ImplementationCpG Transformer is freely available at https://github.com/gdewael/cpg-transformer.

bioinformatics↗