Biology subjects

Smith, M. B.

Publications and source records attributed to Smith, M. B..

13 recordsLinked to original sources

bioRxiv2026

NuclearIDTracker resolves intestinal cell identity and lineage dynamics through nuclear phenotypic signatures

Organoid models have transformed our understanding of intestinal renewal. Fluorescent imaging has been extensively used to identify key cell types and their differentiation pathways, but immunofluorescence provides only static readouts, whereas live imaging requires fluorescent-reporter engineering and is constrained by limited multiplexing and spectral overlap. Here, we introduce NuclearIDTracker, an explainable machine-learning framework that infers cell identity directly from 3D nuclear segmentations. Using a single nuclear marker, NuclearIDTracker accurately classifies intestinal cell types and integrates with single-cell tracking to resolve lineages and reconstruct dynamic state transitions during organoid development. We show that TA-like cells, rather than stem cells, drive early crypt formation and generate enterocyte and Paneth lineages, as well as the stem-cell population, which emerges only later and subsequently replenishes the TA-like compartment. Following stem-cell ablation, crypt regeneration was not driven by a single discrete cell type. Instead, multiple epithelial populations converged on a proliferative regenerative state with a nuclear phenotypic signature that resembled, but remained distinct from, that of homeostatic TA-like cells, and a YAP/TAZ-associated fetal-like transcriptional signature. Thus, nuclear phenotypic signatures resolve cell identity and reveal coordinated epithelial plasticity during crypt regeneration. NuclearIDTracker establishes a non-perturbative tool to quantify cell identity and state dynamics at single-cell resolution, revealing previously inaccessible biological dynamics and expanding the toolkit for studying epithelial homeostasis, regeneration, and disease.

cell biology↗

bioRxiv2026

UDIST: unsupervised disentanglement of shape and texture for multi-scale phenotypic profiling in 2D microscopy

Microscopy-based phenotypic profiling relies increasingly on autonomous, unsupervised feature extraction, yet no existing method explicitly separates shape from texture into dedicated and independent latent subspaces by architectural design. Therefore texture, encoding critical biological information such as protein distribution and intracellular organisation, remains inaccessible as an independent feature domain in standard unsupervised approaches. This represents a fundamental limitation that prevents unbiased phenotypic analysis across biological scales. Here we introduce UDIST (Unsupervised Disentanglement of Shape and Texture), a sequential dual variational autoencoder (VAE) framework that tackles this fundamental limitation by explicitly decoupling shape from texture into independent, non-overlapping latent subspaces at the single-object level. By training two VICReg-regularised VAEs on principal-axis-aligned objects, UDIST separates binary shape from continuous texture information into rotation-invariant feature spaces, enabling separate downstream analysis of both domains. We validated UDIST across biological scales, from nuclei and single cells to patient-derived intestinal organoids, using both fluorescence and brightfield imaging, revealing phenotypic differences previously hidden by morphological variation and enabling the independent analysis of shape and texture in downstream analyses including clustering and similarity measurements. UDIST provides a versatile, label-free, and unsupervised tool for multi-scale phenotypic profiling in high-content microscopy and screening.

bioinformatics↗

bioRxiv2026

Personalized multi-assay profiling of respiratory motile ciliopathies and mRNA therapy

IntroductionImpaired motile cilia function contributes to many respiratory disorders, but therapies targeting this cellular defect are currently lacking. Personalized airway epithelial models combined with quantitative, complementary ciliary assays can pave the way for the development of such therapies. However, existing airway epithelial cultures often show variable ciliogenesis, and ciliary function is frequently assessed using a single assay that does not capture the phenotypic heterogeneity of ciliary dysfunction. Here, we established a personalized, multi-assay in vitro platform using human nasal epithelial cells (HNECs) to assess ciliary function and therapeutic response, using primary ciliary dyskinesia (PCD) as a model disease. MethodsHNECs from 8 healthy individuals and 13 individuals with PCD carrying distinct disease-associated variants were obtained by nasal brushing. Cells were differentiated under optimized conditions, including {gamma}-secretase/Notch and BMP pathway inhibitors and a low liquid-liquid interface, to generate highly ciliated 2D epithelial cultures. Ciliary function was assessed using ciliary beat frequency, bead transport, and apical-out nasal organoid rotation assays. Therapeutic rescue was assessed in HNECs harboring DNAI1 alterations using DNAI1 mRNA-loaded lipid nanoparticles. ResultsOptimized differentiation yielded reproducibly multiciliated HNEC cultures. The multi-assay platform distinguished healthy from PCD-derived HNECs and revealed individual- and genotype-specific patterns of ciliary dysfunction not captured by a single assay. Basolateral administration of DNAI1 mRNA-loaded lipid nanoparticles resulted in partial, dose-dependent recovery of ciliary function in DNAI1-deficient HNECs. ConclusionThis study establishes a standardized, individual-specific multi-assay nasal epithelial platform for functional phenotyping of motile cilia and preclinical evaluation of emerging therapies, with demonstrated utility in PCD.

cell biology↗

bioRxiv2026

A Perfect Soldier: the black soldier fly as a microbial-mediated physiological resilience model .

Understanding the complex interplay between a host, its diet, and its microbiome is crucial for comprehending an organisms health and adaptability. Diet impacts both the host and microbiome, which then influence each other. We used black soldier fly larvae (Hermetia illucens) as a model to investigate this tripartite interaction due to its resilience and bioconversion capabilities. We analyzed life-history traits and metatranscriptomics in larvae fed three diets: carbohydrate-rich, protein-rich, and balanced. Our results showed that dietary macronutrients correlated with shifts in the microbial community and gene expression. The carbohydrate-rich diet, in particular, led to increased microbial diversity and carbohydrate metabolism transcripts. However, this diet also negatively affected larval weight and development, suggesting potential host control over the microbiome. Overall, black soldier fly performance was highest on the balanced diet. This study highlights the black soldier flys resilience and its value as a model for exploring host-diet-microbe interactions. Significance StatementUnderstanding the intricate interplay between an organism, its diet, and its microbiome is fundamental to health and adaptability. This complex tripartite relationship, where dietary macronutrients influence microbial communities and their gene expression, while the host maintains control, is crucial for addressing global challenges from sustainable food systems to personalized medicine. Using the black soldier fly as a resilience model, our metatranscriptomic study reveals how specific dietary shifts impact both host and microbial gene expression, providing mechanistic insights into nutrient utilization and adaptability with broad implications for diverse biological systems, including animals and humans.

systems biology↗

bioRxiv2025

Protein Abundance Inference via Expectation Maximization in Fluorosequencing

1Fluorosequencing produces millions of single-peptide reads, yet a principled strategy for converting these data into quantitative protein abundances has been lacking. We introduce a probabilistic framework that adapts expectation maximization to the fluorosequencing measurement process, estimating relative protein abundances with peptide inference results delivered by previously developed peptide-classification tools. The algorithm iteratively updates protein abundances, maximising the likelihood of the observed reads by obtaining more accurate protein abundance estimations. We first assess performance on simulated five-protein mixtures that reflect realistic labelling and system errors. A simple Python implementation processes one million reads in under ten seconds on a standard work-station and lowers the mean absolute error in relative abundance by more than an order of magnitude compared with a uniform-abundance guess, demonstrating robustness in protein inference for small-scale settings. Scalability is then evaluated with simulations of the complete human proteome (20 642 proteins). Ten million reads are processed in less than four hours on a NVIDIA DGX system using one Tesla V100 GPU, confirming that the method remains tractable at proteome scale. Using error rates characteristic of current fluorosequencing, the algorithm produces marginal improvements in relative abundance accuracy. However, when error rates were artificially lowered, estimation error decreased significantly. This result suggests that improvements in fluorosequencing chemistry could directly translate into substantially more accurate quantitative proteomics with this computational framework. Together, these results establish EM-based inference as a scalable model-driven bridge between peptide-level classification and protein-level quantification in fluorosequencing, laying computational groundwork for high-throughput single-molecule proteomics. Furthermore, the proposed protein inference framework can also be used as a refinement step within other inference methods, enhancing their protein abundance estimates. 2 Author summaryProteins carry out many of the functions inside our cells, but measuring how many copies of each protein are present remains a difficult task. The protein sequencing technology called fluorosequencing observes individual protein fragments one by one, producing millions of short "snapshots" in a single experiment. While this flood of data holds great promise, it is not yet clear how to turn the snapshots into reliable counts of the original proteins. In this study we developed a mathematical procedure, based on a classic "guess-and-improve" strategy called expectation maximization, that bridges this gap. Our program takes the computers best guesses for each fragment and repeatedly refines protein abundances until they best explain all of the measurements. Using realistic computer-generated data, we show that the method is both fast and accurate: it finishes in seconds for small mixtures and in only a few hours for the entire human set of proteins, both with accessible hardware. Because our approach can work with any future improvements in fragment interpretation, it lays the essential groundwork for bringing single-molecule protein sequencing into routine biological and medical use.

bioinformatics↗

bioRxiv2024

Cryo-electron tomography reveals the nanoscale architecture of the actin cortex in cellular blebs

In animal cells, cellular deformations driving cytokinesis, migration, and epithelial constriction are driven by contractile tension in the actomyosin cortex, a thin network of actin and myosin underlying the plasma membrane. Cortical tension results from myosin-generated forces and as such, cortical myosin organization and dynamics have received significant attention. However, recent studies highlight that alongside myosin motor activity, the organization of the cortical actin network is a key regulator of tension. Yet, very little is known about the structural arrangement of cortical actin filaments. This is mostly due to the small thickness and high density of the cortex, which makes the visualization of cortical actin filaments challenging. Here, we use cryo-electron tomography (cryo-ET) to unveil the structural organization of cortical actin. As a model, we use isolated cellular blebs, which assemble an actin cortex comparable to the cortex of entire cells, but are small enough to be amenable to cryo-ET. We find that the bleb actin cortex is mostly composed of short and straight actin filaments. We then characterize cortex structural parameters, including the density of potential cross-linking and membrane attachment points. Our study unveils the nanoscale three-dimensional organization of the cortical actin network in cellular blebs. As such, it provides a quantitative framework for models of cortical tension generation, and will help understanding the nanoscale basis of cell surface contractions.

biophysics↗

bioRxiv2023

Synergistic action of actin binding proteins regulate actin network organization and cell shape

Animal cell shape is largely determined by the organization of the actin cytoskeleton. Spread shapes result from a balance between protrusive actin networks and contractile stress fibers, while rounded shapes are supported by a contractile actomyosin cortex. The assembly and regulation of distinct types of actin networks have been extensively studied, yet, what determines which networks dominate in a given cell remains unclear. In this Brief Report, we explore the molecular regulation of overall actin organization and resulting cell shape. We use our recently published comparison of the F-actin interactome in spread interphase and rounded mitotic cells to establish a list of candidate regulators of actin networks in spread cells. Utilizing micropatterning and automated image analysis we quantitatively analyze how these candidates affect actin organization. Out of our initial 16 candidates, we identify subsets of proteins promoting stress fibers or regulating their arrangement. Interestingly, no single regulator depletion caused significant cell shape change. However, perturbing two hits simultaneously, supervillin and myosin II, led to stress fiber disassembly and cell rounding. Overall, our systematic investigation shows that actin networks are robust to perturbations, and identifies regulatory modules controlling overall actin organization and resulting cell shape.

cell biology↗

bioRxiv2023

Estimating error rates for single molecule protein sequencing experiments

The practical application of new single molecule protein sequencing (SMPS) technologies requires accurate estimates of their associated sequencing error rates. Here, we describe the development and application of two distinct parameter estimation methods for analyzing SMPS reads produced by fluorosequencing. A Hidden Markov Model (HMM) based approach, extends whatprot, where we previously used HMMs for SMPS peptide-read matching. This extension offers a principled approach for estimating key parameters for fluorosequencing experiments, including missed amino acid cleavages, dye loss, and peptide detachment. Specifically, we adapted the Baum-Welch algorithm, a standard technique to estimate transition probabilities for an HMM using expectation maximization, but modified here to estimate a small number of parameter values directly rather than estimating every transition probability independently, which should help prevent overfitting. We demonstrate a high degree of accuracy on simulated data, but on experimental datasets, we observed that the model needed to be augmented with an additional error type, N-terminal blocking. This, in combination with data pre-processing, results in reasonable parameterizations of experimental datasets that agree with controlled experimental perturbations. A second independent implementation using a hybrid of DIRECT and Powells method to reduce the root mean squared error (RMSE) between simulations and the real dataset was also developed. We compare these methods on both simulated and real data, finding that our Baum-Welch based approach outperforms DIRECT and Powells method by most, but not all, criteria. Although some discrepancies between the results exist, we also find that both approaches provide similar error rate estimates from experimental single molecule fluorosequencing datasets.

bioinformatics↗

bioRxiv2023

Signalling-dependent refinement of cell fate choice during tissue remodelling

How biological form emerges from cell fate decisions and tissue remodelling is a fundamental question in development biology. However, an understanding of how these processes operate side-by-side to set precise and robust patterns is largely missing. Here, we investigate this interplay during the process of vein refinement in the Drosophila pupal wing. By following reporters of signalling activity dynamically, together with tissue flows, we show that longitudinal vein refinement arises from a combination of local tissue deformation and cell fate adjustments controlled by a signalling network involving Notch, Dpp, and EGFR. Perturbing large-scale convergence and extension tissue flows does not affect vein refinement, showing that pre-patterned vein domains are able to intrinsically refine to the correct width. A minimal biophysical description taking into account key signalling interactions recapitulates the intrinsic tissue ability to establish a thin, regular vein independently of large-scale tissue flows. Supporting this prediction, artificial proveins optogenetically generated orthogonal to the axis of wing elongation refine against large-scale flows. Overall, we find that signalling-mediated updating of cell fate is a key contributor to reproducible patterning.

developmental biology↗

bioRxiv2023

Active mesh and neural network pipeline for cell aggregate segmentation

Segmenting cells within cellular aggregates in 3D is a growing challenge in cell biology, due to improvements in capacity and accuracy of microscopy techniques. Here we describe a pipeline to segment images of cell aggregates in 3D. The pipeline combines neural network segmentations with active meshes. We apply our segmentation method to cultured mouse mammary duct organoids imaged over 24 hours with oblique plane microscopy, a high-throughput light-sheet fluorescence microscopy technique. We show that our method can also be applied to images of mouse embryonic stem cells imaged with a spinning disc microscope. We segment individual cells based on nuclei and cell membrane fluorescent markers, and track cells over time. We describe metrics to quantify the quality of the automated segmentation. Our segmentation pipeline involves a Fiji plugin which implement active meshes deformation and allows a user to create training data, automatically obtain segmentation meshes from original image data or neural network prediction, and manually curate segmentation data to identify and correct mistakes. Our active meshes-based approach facilitates segmentation postprocessing, correction, and integration with neural network prediction. Statement of significanceIn vitro culture of organ-like structures derived from stem cells, so-called organoids, allows to image tissue morphogenetic processes with high temporal and spatial resolution. Three-dimensional segmentation of cell shape in timelapse movies of these developing organoids is however a significant challenge. In this work, we propose an image analysis pipeline for cell aggregates that combines deep learning with active contour segmentations. This combination offers a flexible and efficient way to segment three-dimensional cell images, which we illustrate with by segmenting datasets of growing mammary gland organoids and mouse embryonic stem cells.

bioinformatics↗

bioRxiv2022

Amino acid sequence assignment from single molecule peptide sequencing data using a two-stage classifier

We present a machine learning-based interpretive framework (whatprot) for analyzing single molecule protein sequencing data produced by fluorosequencing, a recently developed proteomics technology that determines sparse amino acid sequences for many individual peptide molecules in a highly parallelized fashion [1] [2]. Whatprot uses Hidden Markov Models (HMMs) to represent the states of each peptide undergoing the various chemical processes during fluorosequencing, and applies these in a Bayesian classifier, in combination with pre-filtering by a k-Nearest Neighbors (kNN) classifier trained on large volumes of simulated fluorosequencing data. We have found that by combining the HMM based Bayesian classifier with the kNN pre-filter, we are able to retain the benefits of both, achieving both tractable runtimes and acceptable precision and recall for identifying peptides and their parent proteins from complex mixtures, outperforming the capabilities of either classifier on its own. Whatprots hybrid kNN-HMM approach enables the efficient interpretation of fluorosequencing data using a full proteome reference database and should now also enable improved sequencing error rate estimates.

bioinformatics↗

bioRxiv2020

ECM remodeling and spatial cell cycle coordination determine tissue growth kinetics

During development, multicellular organisms undergo stereotypical patterns of tissue growth to yield organs of highly reproducible sizes and shapes. How this process is orchestrated remains unclear. Analysis of the temporal dynamics of tissue growth in the Drosophila abdomen reveals that cell cycle times are spatially correlated and that growth termination occurs through the rapid emergence of a population of arrested cells rather than a gradual slowing down of cell cycle time. Reduction in apical tension associated with tissue crowding has been proposed as a developmental growth termination mechanism. Surprisingly, we find that growth arrest in the abdomen occurs while apical tension increases, showing that in this tissue a reduction in tension does not underlie the mechanism of growth arrest. However, remodeling of the extracellular matrix is necessary for tissue expansion. Thus, changes in the tissue microenvironment, and a rapid exit from proliferation, control the formation of the adult Drosophila abdomen.

developmental biology↗

bioRxiv2019

Dynamic colonization of microbes and their functions after fecal microbiota transplantation for inflammatory bowel disease

For fecal microbiota transplantation (FMT) to be successful in complex immune diseases like inflammatory bowel disease (IBD), it is assumed that therapeutic microbes and their beneficial functions and immune interactions must colonize the recipient and persist in sufficient quantity and for a long enough period of time to result in a clinical benefit. But few studies have comprehensively profiled the colonization and persistence of transferred microbes along with the transfer of their microbial and immune functions. Using 16S, metagenomic, and immunoglobulin A (IgA) sequencing, we analyzed hundreds of longitudinal microbiome samples from a randomized controlled trial of 12 patients with ulcerative colitis who received fecal transplant or placebo for 12 weeks. We uncovered a range of competitive dynamics among donor and patient strains, showing that persistence of transferred microbes is far from static. Indeed, one patient experienced dramatic loss of donor bacteria 10 weeks into the trial, coinciding with a bloom of pathogenic bacteria and worsening clinical symptoms. We similarly evaluated transfer of microbial functions, including desired ones like butyrate production and unintended ones like antibiotic resistance. By profiling bacteria coated with IgA, we identified IgA-coated bacteria associated with inflammation, and we found that microbial interactions with the host immune system can be transferred across people. This transfer of immune function is likely critical for gut microbiome therapeutics for immune-related diseases. Our findings elucidate the colonization dynamics of gut microbes as well as their functions in the context of FMT to treat a complex disease--information that may provide a critical foundation for the development of more-targeted therapeutics.

microbiology↗