Biology subjects

Bonneau, R.

Publications and source records attributed to Bonneau, R..

At least 19 recordsLinked to original sources

bioRxiv2026

MIMOSA: Guiding De Novo Antibody Design with Natural Interaction Fingerprints

De novo antibody design promises to transform therapeutic antibody and nanobody discovery. Yet, success remains inconsistent across targets, with generative models often yielding no experimentally validated binders. Here we introduce MIMOSA (MIMic-Oriented Structural Antibody generation), a model-agnostic, inference-time framework that constrains diffusion-based antibody design pipelines to reproduce the interaction geometry and chemistry of a target's known cognate binder interface. Rather than searching for a productive interface from scratch, MIMOSA constrains generation around residue identities and geometries already known to support binding, while allowing the generative model to complete the surrounding antibody interface. We test MIMOSA on different interface topologies: contiguous motifs that fit within a single CDR loop and spatially dispersed hotspots distributed across the paratope. Applied without retraining to two architecturally distinct generators (RFAntibody and BoltzGen), this framework delivers experimentally validated binders-achieving sub-micromolar to single-digit nanomolar affinities-at 20-26% hit rates on three targets on which unconstrained methods produce zero or near-zero binders: KEAP1, uPA, and IL-8. By turning any target with a structurally characterised cognate binder into an accessible de novo design problem, this framework broadens the practical reach of generative antibody design to targets where unconstrained approaches currently fail.

bioinformatics↗

bioRxiv2025

Moving from Association to Causation: Instrumental factor models for causal inference in high-dimensional multi-omics data

Distinguishing true causal relationships from mere association is a core challenge in bioinformatics and science in general. Causal inference, in particular in "-omics" studies, is hindered by high dimensionality, correlated measurements, and pervasive endogeneity arising from unmeasured confounding and complex interventions. We introduce Factor IV, a supervised instrumental variable (IV) framework for identifying causal effects in high-dimensional biological systems. FactorIV constructs low-dimensional instrumental factors via a sparse low-rank decomposition of the instrument-exposure map, yielding an identifiable first stage even when either/both endogenous and instrumental variables are high-dimensional. Under standard IV assumptions, these factors remain orthogonal to outcome noise while capturing coordinated, perturbation-driven biological variation. The framework supports generalized first-stage models, including Gaussian, Bernoulli and negative binomial likelihoods. Simulation studies demonstrate accurate recovery of factor-level and feature-level causal effects under linear and generalized settings, with robustness to correlated errors and hidden confounding. To illustrate biological discovery, we applied FactorIV to a mouse hepatocellular carcinoma model and the DIABIMMUNE infant cohort. We demonstrate how FactorIV uncovers mechanistically interpretable causal structure beyond association-based analyses.

bioinformatics↗

bioRxiv2025

CDR Conformation Aware Antibody Sequence Design with ConformAb

Antibody lead optimization methods seek to enhance a lead candidates therapeutic properties through targeted sequence mutation. However, the mutations introduced during this process can inadvertently induce structural changes that disrupt binding, particularly by altering CDR loop conformations which destabilize CDR-target binding interactions. To address this, we present ConformAb, a guided discrete diffusion model for designing antibody sequences that explicitly conform to the CDR canonical conformation of the lead. With seed sequences as the starting point, we show that our method is capable of generating 3- 5x better binders that conform to the same CDR backbone structure in a one-shot design setup, and in some cases, binders with better affinity than seed repertoire picks without any target-specific data included in training. Across the targets tested, ConformAb demonstrated binding rates ranging from 15 -60% in wet lab experiments, a result obtained using fewer than 100 designs for each target. ConformAb offers a unique one-shot approach for antibody lead optimization in data-scarce scenarios where experimental/repertoire data cannot be leveraged for model training.

biophysics↗

bioRxiv2025

Property Enhancer - a data efficient multi-objective approach for functional antibody optimization

In-silico antibody lead optimization remains challenging due to scarce high-quality data, costly experimental validation, and the need to jointly optimize multiple developability properties. Discovery workflows often rely on high-throughput phage, ribosome or yeast display experiments, which yield large but noisy datasets; as leads emerge, strategies shift to low-throughput assays which are precise, yet unscalable. Deep-learning and language-model approaches are hindered by such limited, unreliable measurements. We introduce Property Enhancer (PropEn), a data-efficient framework for low-data, heterogeneous regimes that can simultaneously optimize multiple antibody properties. PropEn proposes a matching-based augmentation that expands the training data with sequence pairs differing by only a few mutations; within each pair the second sequence improves the target value, providing an implicit optimization signal. Extensive in silico and in vitro tests show 10-39x affinity gains across four targets and nine leads, and enable joint multi-property optimization, positioning PropEn as a scalable, general solution.

molecular biology↗

bioRxiv2025

Dynamic Gene Regulatory Network Inference with Interpretable, Biophysically-Motivated Neural ODEs

Gene Regulatory Networks (GRNs) are complex dynamical systems that modulate gene expression and drive transitions between phenotypic cell states. Determining these networks is crucial in understanding how gene dysregulation can lead to phenotypic variation and perturbation responses. We present a novel biophysically-motivated neural ordinary differential equation (ODE) model framework with a biologically interpretable deep learning architecture that leverages dynamic single-cell data: in-CAHOOTTS (gene regulatory network Inference with Context Aware Hybrid neural-ODEs on Transcriptional Time-series Systems). Our approach combines accurate prediction with mechanistic interpretability, decomposing gene expression dynamics into fundamental biophysical processes (mRNA transcription and degradation) while inferring regulatory network structure. We validate in-CAHOOTTS by learning the Saccharomyces cerevisiae response to rapamycin treatment, and learning the dynamic cell cycle progression. The trained model accurately predicts gene expression trajectories and identifies biologically relevant rapamycin response regulators. The framework achieves stable long-term predictions with time frames extending 30 times beyond training data, maintaining realistic oscillatory dynamics for over 40 hours in cell cycle modeling, demonstrating that neural ODEs can capture true biological attractors rather than merely fitting data. We argue that interpretable neural ODEs can successfully model complex biological dynamics while revealing mechanistic insights essential for understanding living systems. By bridging extensible machine learning frameworks and mechanistic biology, in-CAHOOTTS represents a significant advance toward deep learning systems that both predict biological outcomes and explain the mechanisms driving them.

systems biology↗

bioRxiv2025

PlasmoFP: leveraging deep learning to predict protein function of uncharacterized proteins across the malaria parasite genus

The first malaria parasite Plasmodium falciparum genome published in 2002 jump-started functional studies, but a large fraction of all predicted proteins remains partially annotated and of unknown function. Here, we introduce Plasmodium Function Predictor (PlasmoFP), deep learning models designed specifically for species of genus Plasmodium. Innovatively, PlasmoFP models are trained on structure-function relationships of the phylogenetically relevant SAR (Stramenopiles, Alveolate, and Rhizarians) supergroup proteins, addressing challenges to annotating Plasmodium proteins due to their low sequence similarity to well-characterized model organism proteins. PlasmoFP models estimate epistemic uncertainty, control false discovery rates in model predictions, and are validated using proteins with manually curated GO terms and experimentally characterized proteins. Integrating PlasmoFP predictions with current protein annotations, we reduced the proportion of unannotated proteins without Gene Ontology terms from 15-59% to 3-28% across 19 Plasmodium species, and improved the proportion of fully annotated proteins from 7-42% to 36-68%. PlasmoFP predictions advance Plasmodium basic research, an important component of global malaria R&D.

bioinformatics↗

bioRxiv2025

Deep Evolutionary Fitness Inference for Variant Nomination from Directed Evolution

Iterative screening techniques, such as directed evolution, enable high-throughput affinity maturation to optimize binders to molecular interfaces. However, the decision problem of selecting variants from rich, evolved populations to enter low-throughput follow-up methods remains a significant bottleneck. Here, we present evolutionary fitness inference (EVFI) and DeepEVFI, two machine learning methods that model directed evolution from time-series sequencing data, and infer fitness, a variants ability to enrich under selection pressure. Our methods flexibly handle mutation mechanisms and starting populations that may be partially unknown - settings relevant to drug discovery - and achieve strong performance on a diverse set of experimental data. We conducted two experimental directed evolution campaigns, using antibodies and macrocyclic peptides libraries to identify and optimize binders to therapeutically relevant targets. EVFI and DeepEVFI identified tighter binders that were missed by human experts using conventional frequency-based approaches, including "rising stars" with low frequency. Beyond initial hit discovery, EVFI and Deep-EVFI enables labeling large-scale sequence-fitness datasets and identifying variants of initial binders with diverse properties.

bioinformatics↗

bioRxiv2025

GLM-Prior: a nucleotide transformer model reveals prior knowledge as the driver of GRN inference performance

Gene regulatory network (GRN) inference depends on high-quality prior knowledge, yet curated priors are often incomplete or unavailable across species and cell types. We present GLM-Prior, a genomic language model fine-tuned to predict transcription factor (TF)-target gene interactions from nucleotide sequence, and benchmark it as a sequence-derived strategy for constructing TF-gene prior matrices. We integrate GLM-Prior with PMF-GRN, a probabilistic matrix factorization model, to create a dual-stage pipeline that combines sequence-derived priors with single-cell gene expression data for prior-conditioned GRN inference. Across six human, mouse, and yeast cell-line contexts, GLM-Prior performance scales with positive label abundance and TF coverage in training data, and shows abovechance agreement with independent reference networks in well-annotated mammalian contexts. We evaluate single-species, species-transfer, and multi-species training paradigms, finding that GLM-Prior can construct informative priors across related mammalian species, while transfer to yeast remains near chance. Comparisons with accessibility-based priors across multiple GRN inference methods show that GLM-Prior achieves the highest prior performance in four of five mammalian cell lines. Under these benchmarks, prior quality largely constrains achievable GRN inference performance, with expression-based inference providing prior-dependent refinement. Together, these results position GLM-Prior as a benchmarked workflow for transferable, sequence-derived prior construction in systems where matched experimental assays are unavailable.

genomics↗

bioRxiv2025

Genome-wide expression gradient estimation based on local pseudotime in single cell RNA sequencing

Single cell genomics measures the internal state of individual cells and makes it possible to describe biological phenomena such as cell type heterogeneity, developmental progression, and dynamic differentiation using computational methods. A prominent approach to facilitate downstream analysis, for large collection of cells, is to connect individual cells by a k-Nearest Neighbor Graph (kNN-G). The kNN-G is among other things used by methods that derive pseudotime, a temporal ordering of cells. However, pseudotime methods require knowledge of initial states of the trajectories created. Alternatives have been invented that estimate gene rate of transcription e.g. by RNA velocity but these methods dont generalize to all individual genes for estimating transcriptional rate of change along pseudotime. While these methods can derive pseudo progression de novo by extrapolating transcription velocity vectors they are limited to subsets of genes where intronic reads are captured with sufficient magnitude. Here we present pseudovelocity, an alternative method for calculating a local rate of change in transcription based on the kNN-G and diffusion-based pseudotime, to derive RNA velocity estimates for individual genes, and demonstrate how it improves the gene-based estimation of RNA velocity for downstream analysis. This package can be accessed here: https://gitlab.com/Xparx/pseudovelocity.

bioinformatics↗

bioRxiv2025

DyAb: sequence-based antibody design and property prediction in a low-data regime

Protein therapeutic design and property prediction are frequently hampered by data scarcity. Here we propose a new model, DyAb, that addresses these issues by leveraging a pair-wise representation to predict differences in protein properties, rather than absolute values. DyAb is built on top of a pre-trained protein language model and achieves a Spearman rank correlation of up to 0.85 on binding affinity prediction across molecules targeting three different antigens (EGFR, IL-6, and an internal target), given as few as 100 training data. We employ DyAb in two design contexts: as a ranking model to score combinations of known mutations, and combined with a genetic algorithm to generate new sequences. Our method consistently generates novel antibody candidates with high binding rates, including designs that improve on the binding affinity of the lead molecule by more than ten-fold. DyAb represents a powerful tool for engineering therapeutic protein properties in low data regimes common in early-stage drug development.

bioengineering↗

bioRxiv2025

deepNGS Navigator: Exploring antibody NGS datasets using deep contrastive learning

High-throughput sequencing uncovers how B-cells adapt and evolve in response to antigens by generating B-cell receptor (BCR) sequences at an unprecedented scale. As BCR datasets grow to be millions of sequences, using efficient computational methods becomes crucial for analyzing and understanding complex patterns within the data. One important aspect of antibody sequence analysis is detecting clonal families or clusters of related sequences, whether they come from immunization, synthetic libraries or even ML-generated datasets. Such analysis helps us to understand how sequences are evolutionarily connected, and how they might have been selected or evolved. Here we introduce deepNGS Navigator, a computational tool that leverages language models and contrastive learning to transform large datasets of antibody sequences into 2D representations. The resulting 2D maps offer an intuitive visualization of overall diversity of input datasets, which can be clustered based on the sequence distances and their densities across the map. Beyond grouping related sequences, the 2D maps can also point to evolutionary trajectories and capture mutational patterns among closely related sequences. By analyzing properties like charge, hydrophobicity, number of sequence neighbors, and read counts, the maps highlight which clusters are most promising for further investigation while also detecting anomalies or noisy sequences with higher risk. We demonstrate deepNGS Navigators utilities on various datasets, including: 1) a synthetic library from a yeast display targeting HER2, 2) a machine learning-generated dataset with a hierarchical tree structure, 3) NGS sequences from a llama immunized against COVID RBD, 4) human naive and memory B-cell sequences, and 5) an insilico dataset simulating B-cell clonal lineages.

bioinformatics↗

bioRxiv2024

Comparative genomics of the sexually transmitted parasite Trichomonas vaginalis reveals relaxed and convergent evolution and genes involved in spillover from birds to humans

Trichomonas vaginalis is the causative agent of the venereal disease trichomoniasis which infects men and women globally and is associated with serious outcomes during pregnancy and cancers of the human reproductive tract. Trichomonads parasitize a range of hosts in addition to humans including birds, livestock, and domesticated animals. Recent genetic analysis of trichomonads recovered from columbid birds has provided evidence that these parasite species undergo frequent host-switching, and that a current epoch spillover event from columbids likely gave rise to T. vaginalis in humans. We undertook a comparative evolutionary genomics study of seven trichomonad species, generating chromosome-scale reference genomes for T. vaginalis and its avian sister species Trichomonas stableri, and assemblies of five other species that infect birds and mammals. Human-infecting trichomonad lineages have undergone recent and convergent genome size expansions compared to their avian sister species, and the major contributor to their increased genome size is increased repeat expansions, especially multicopy gene families and transposable elements, with genetic drift likely a driver due to relaxed selection. Trichomonads have independently host-switched twice from birds to humans, and genes implicated in the transition to the human host include those associated with host tissue adherence and phagocytosis, extracellular vesicles, and CAZyme virulence factors.

microbiology↗

bioRxiv2024

Generating All-Atom Protein Structure from Sequence-Only Training Data

AbstractO_ST_ABSPurposeC_ST_ABSDesigning proteins with atomic-level functional control remains a central challenge in de novo design, exacerbated by limitations in structural data availability. MethodsWe introduce PLAID (Protein Latent Induced Diffusion), a generative model that efficiently co-generates discrete sequence and all-atom structure by sampling directly in the shared sequence-structure latent of a pretrained sequence-to-structure predictor. Unlike existing de novo generative models, PLAID trains its diffusion model exclusively on sequences, expanding effective training corpora by 2-4 orders of magnitude relative to structural databases. Using classifier-free guidance, PLAID supports controllable generation based on function (Gene Ontology) and organism keywords. ResultsIn silico, PLAID can unconditionally generate all-atom structures without explicit structural supervision during diffusion model training. Function conditioned proteins can recapitulates catalytic side-chain positions for residues at non-adjacent positions, and transmembrane proteins with expected hydrophobicity patterns and predicted topologies. We further experimentally validate that PLAID can be prompted to generate heme binding proteins with high sequence novelty. ConclusionOverall, PLAID unifies sequence-scale training with atomic-level generation, enabling more precise functional control in protein design. Code and weights: github.com/amyxlu/plaid.

bioinformatics↗

bioRxiv2024

Tokenized and Continuous Embedding Compressions of Protein Sequence and Structure

Existing protein machine learning representations typically model either the sequence or structure distribution, with the other modality implicit. The latent space of sequence-to-structure prediction models such as ESMFold represents the joint distribution of sequence and structure; however, we find these embeddings to exhibit massive activations, whereby some channels have values 3000x higher than others, regardless of the input. Further, on continuous compression schemes, ESMFold embeddings can be reduced by a factor of 128x along the channel and 8x along the length, while retaining structure information at <2[A] scale accuracy, and performing competitively on protein function and localization benchmarks. On discrete compression schemes, we construct a tokenized all-atom structure vocabulary that retains high reconstruction accuracy, thus introducing a tokenized representation of all-atom structure that can be obtained from sequence alone. We term this series of embeddings as CHEAP (Compressed Hourglass Embedding Adaptations of Proteins) embeddings, obtained via the HPCT (Hourglass Protein Compression Transformer) architecture. CHEAP is a compact representation of both protein structure and sequence, sheds light on information content asymmetries between sequence and structure, democratizes representations captured by large models, and is designed to have flexible downstream applications such as generation, search, and prediction.

bioinformatics↗

bioRxiv2024

Cramming Protein Language Model Training in 24 GPU Hours

Protein language models (pLMs) are ubiquitous across biological machine learning research, but state-of-the-art models like ESM2 take hundreds of thousands of GPU hours to pre-train on the vast protein universe. Resource requirements for scaling up pLMs prevent fundamental investigations into how optimal modeling choices might differ from those used in natural language. Here, we define a "cramming" challenge for pLMs and train performant models in 24 hours on a single GPU. By re-examining many aspects of pLM training, we are able to train a 67 million parameter model in a single day that achieves comparable performance on downstream protein fitness landscape inference tasks to ESM-3B, a model trained for over 15, 000x more GPU hours than ours. We open source our library1 for training and inference, LBSTER: Language models for Biological Sequence Transformation and Evolutionary Representation.

bioengineering↗

bioRxiv2024

Tissue and cellular spatiotemporal dynamics in colon aging

Tissue structure and molecular circuitry in the colon can be profoundly impacted by systemic age-related effects, but many of the underlying molecular cues remain unclear. Here, we built a cellular and spatial atlas of the colon across three anatomical regions and 11 age groups, encompassing [~]1,500 mouse gut tissues profiled by spatial transcriptomics and [~]400,000 single nucleus RNA-seq profiles. We developed a new computational framework, cSplotch, which learns a hierarchical Bayesian model of spatially resolved cellular expression associated with age, tissue region, and sex, by leveraging histological features to share information across tissue samples and data modalities. Using this model, we identified cellular and molecular gradients along the adult colonic tract and across the main crypt axis, and multicellular programs associated with aging in the large intestine. Our multi-modal framework for the investigation of cell and tissue organization can aid in the understanding of cellular roles in tissue-level pathology.

systems biology↗

bioRxiv2024

Meta-analysis of the human gut microbiome uncovers shared and distinct microbial signatures between diseases

Microbiome studies have revealed gut microbiotas potential impact on complex diseases. However, many studies often focus on one disease per cohort. We developed a meta-analysis workflow for gut microbiome profiles and analyzed shotgun metagenomic data covering 11 diseases. Using interpretable machine learning and differential abundance analysis, our findings reinforce the generalization of binary classifiers for Crohns disease (CD) and colorectal cancer (CRC) to hold-out cohorts and highlight the key microbes driving these classifications. We identified high microbial similarity in disease pairs like CD vs ulcerative colitis (UC), CD vs CRC, Parkinsons disease vs type 2 diabetes (T2D), and schizophrenia vs T2D. We also found strong inverse correlations in Alzheimers disease vs CD and UC. These findings detected by our pipeline provide valuable insights into these diseases. IMPORTANCEAssessing disease similarity is an essential initial step preceding disease-based approach for drug repositioning. Our study provides a modest first step in underscoring the potential of integrating microbiome insights into the disease similarity assessment. Recent microbiome research has predominantly focused on analyzing individual disease to understand its unique characteristics, which by design excludes comorbidities individuals. We analyzed shotgun metagenomic data from existing studies and identified previously unknown similarities between diseases. Our research represents a pioneering effort that utilize both interpretable machine learning and differential abundance analysis to assess microbial similarity between diseases.

microbiology↗

bioRxiv2023

Learning sequence, structure, and function representations of proteins with language models

The sequence-structure-function relationships that ultimately generate the diversity of extant observed proteins is complex, as proteins bridge the gap between multiple informational and physical scales involved in nearly all cellular processes. One limitation of existing protein annotation databases such as UniProt is that less than 1% of proteins have experimentally verified functions, and computational methods are needed to fill in the missing information. Here, we demonstrate that a multi-aspect framework based on protein language models can learn sequence-structure-function representations of amino acid sequences, and can provide the foundation for sensitive sequence-structure-function aware protein sequence search and annotation. Based on this model, we introduce a multi-aspect information retrieval system for proteins, Protein-Vec, covering sequence, structure, and function aspects, that enables computational protein annotation and function prediction at tree-of-life scales.

bioinformatics↗