Biology subjects

Moses, A.

Publications and source records attributed to Moses, A..

12 recordsLinked to original sources

bioRxiv2026

Projectomic Organization of the Serotonin System of the Mouse Brain

The serotonin system innervates nearly the entire brain to support diverse functions, and its dysfunction is implicated in multiple psychiatric disorders. Although recent studies suggested that this anatomically diffuse system supports differentiated rather than uniform modulation, its overall organization is unclear. Here, using systematic whole-brain axon tracing and integrated analyses, we show that the serotonin projectome is organized by functional relatedness rather than physical proximity of its targets. Dorsal and median raphe serotonin neurons partition into five projectomic groups, each preferentially innervating the hippocampus, basal ganglia, cortex, medial interbrain, or brainstem/lateral thalamus, with within-group structure ranging from discrete subgroups to continuous variation. Spatial transcriptomic analyses reveal that each projectomic group corresponds to a distinct combination of transcriptomic clusters, with transcriptome and somatic position jointly predicting projectomic identity. Projection-specific serotonin depletion produces dissociable behavioral phenotypes. Together, projectomic identity emerges as a central axis linking axon collateralization, molecular diversity, and behavioral function.

neuroscience↗

bioRxiv2026

PRRC2A, PRRC2B and PRRC2C are Stress Granule Proteins that Promote Translation Through Association with the eIF3 complex

Regulation of mRNA translation is essential for cellular homeostasis, and its dysregulation contributes to cancer, neurodegeneration, and developmental disorders. Stress granules are cytosolic condensates that form during stress-induced translation arrest and are enriched in mRNAs, translation factors, and RNA-binding proteins, but how stress granule proteins modulate translation remains poorly understood. Here, we identify the stress granule components Proline-Rich Coiled-Coil A, B, and C (PRRC2 proteins) as translation regulators. PRRC2 proteins are large, intrinsically disordered paralogs conserved across jawed vertebrates. Functional proteomics revealed that all PRRC2 proteins associate with the 48S translation initiation complex (PIC), whereas PRRC2B additionally interacts with nuclear proteins. Under stress, the proximal interaction network of PRRC2 proteins undergoes dynamic remodeling, including increased interactions with the stress granule scaffold G3BP1. Genetic perturbation shows that the PRRC2 proteins influence stress granule assembly in a context-specific manner, and are collectively required for cell growth in basal conditions due to their essential role in translation. Cells with reduced PRRC2 proteins exhibit a significant reduction in the abundance of more than half of the proteome, with a bias toward translational targets of eIF3d and eIF4G2. Interaction domain mapping and AlphaFold3 modeling revealed that an helix within the putative coiled-coil domain of PRRC2C mediates interactions with the eIF3 core complex. This modeling places the PRRC2C helix in a previously unassigned region of a published cryo-EM density map, validating the protein interaction and the mechanistic role of PRRC2C in translation control. Together, these findings establish PRRC2 proteins as components of the translation initiation machinery that regulate translation through their interactions with the eIF3 complex and other components of the 48S PIC factors, providing a direct mechanistic link between stress granule proteins and translational control.

cell biology↗

bioRxiv2026

Predicting evolutionary rate as a pretraining task improves genome language model representations

Genome language models (gLM) have the potential to further understanding of regulatory genomics without requiring labeled data. Most gLMs are pretrained using sequence reconstruction tasks inspired by natural language processing, but recent studies have shown that these gLMs often fail to capture biological signal. To overcome this, we introduce pretraining tasks that predict the rate of evolution. These tasks are designed so that they can be composed with sequence reconstruction, enabling a controlled comparison of predicting sequence only, evolutionary rate only, or both. To address gaps in existing evaluations, we developed a suite of biologically grounded benchmarks. Across these tasks, and for established variant effect prediction benchmarks, models pretrained on both sequence and evolutionary rate outperform those trained on sequence alone, and training on evolutionary rate can make the even the relatively small models in our work competitive with much larger existing gLMs for some tasks. These results establish evolution as a key training target for genome-scale models.

genomics↗

bioRxiv2025

Prevalence of zoonotic hepatic nematode varies with small mammal community diversity across a heterogenous landscape in Eastern Uganda

Identifying key drivers of pathogen infection prevalence and intensity in wildlife is essential to understand disease dispersal and transmission. Calodium hepatica (syn. Capillaria hepatica) is a generalist nematode that infects liver parenchyma of mammals worldwide and is capable of human infections. Prevalence ranges from 0-100% in wildlife, often varying across small geographic areas, making it an ideal parasite for understanding ecological drivers of variation. Here, we quantify prevalence of Calodium hepatica and present initial surveys of synanthropic small mammals in four villages representing differing land cover. Cross-sectional rodent trapping was conducted within and around households over consecutive dry seasons in Eastern Uganda. 18s rRNA gene of C. hepatica was amplified and a sub-set of PCR products sequenced to confirm presence of C. hepatica. Landscape structural diversity was classified by tree crown density and mean canopy height derived from 30m LiDAR data within a 0.5km buffer. Multivariable binomial generalised linear models were fit to C. hepatica prevalence. C. hepatica infection was common (overall 34.5%, CI95% 27.9-41.0) and found in rodent and shrew species inside and outside residences. We observe village-level differences in prevalence (18.2%-75.0%), with higher C. hepatica prevalence associated with higher relative proportion of Rattus rattus to other species (aOR=2.22 CI95% 1.30-3.85). Host diversity appears to be protective against parasite prevalence. Differences in molecular and macroscopy identification highlight challenges in diagnosis and a need for more specialized molecular tools. Further investigation is required to understand individual host and community variation in pathogen infection intensity and implications for zoonotic risk. Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=127 SRC="FIGDIR/small/674808v1_ufig1.gif" ALT="Figure 1"> View larger version (36K): org.highwire.dtl.DTLVardef@e9cfdaorg.highwire.dtl.DTLVardef@4d0f6dorg.highwire.dtl.DTLVardef@12fe43dorg.highwire.dtl.DTLVardef@1318d83_HPS_FORMAT_FIGEXP M_FIG C_FIG Key FindingsO_LIZoonotic reservoirs, including Rattus rattus, Mastomys erythroleucus and Crocidura olivieri, are abundant in human-modified habitats and show elevated levels of household incursion C_LIO_LIHigh prevalence of Calodium hepatica, a nematode endoparasite, was identified in a range of small mammal species C_LIO_LIHigher proportion of native small mammal species relative to Rattus rattus appears protective against zoonotic pathogen load, with higher village-level C. hepatica prevalence in ecologically depleted sites C_LIO_LISmall-holder agriculture may provide a dilution effect through secondary wildlife support and small mammal competition, while dense village settings potentiate C. hepatica prevalence C_LI

ecology↗

bioRxiv2025

Representation Learning Methods for Single-Cell Microscopy are Confounded by Background Cells

Deep learning models are widely used to extract feature representations from microscopy images. While these models are used for single-cell analyses, such as studying single-cell heterogeneity, they typically operate on image crops centered on individual cells with background information present, such as other cells, and it remains unclear to what extent the conclusions of single-cell analyses may be altered by this. In this paper, we introduce a novel evaluation framework that directly tests the robustness of crop-based models to background information. We create synthetic single-cell crops where the center cells localization is fixed and the background is swapped--e.g., with backgrounds from other protein localizations. We measure how different backgrounds affect localization classification performance using model-extracted features. Applying this framework to three leading models for single-cell microscopy for analyzing yeast protein localization, we find that all lack robustness to background cells. Localization classification accuracy drops by up to 15.8% when background cells differ in localization from the center cell compared to when the localization is the same. We further show that this lack of robustness can affect downstream biological analyses, such as the task of estimating proportions of cells for proteins with single-cell heterogeneity in localization. Ultimately, our framework provides a concrete way to evaluate single-cell model robustness to background information and highlights the importance of learning background-invariant features for reliable single-cell analysis.1

bioinformatics↗

bioRxiv2025

Interpreting Attention Mechanisms in Genomic Transformer Models: A Framework for Biological Insights

Transformer models have shown strong performance on biological sequence prediction tasks, but the interpretability of their internal mechanisms remains underexplored. Given their application in biomedical research, understanding the mechanisms behind these models predictions is crucial for their widespread adoption. We introduce a method to interpret attention heads in genomic transformers by correlating per-token attention scores with curated biological annotations, and we use GPT-4 to summarize each heads focus. Applying this to DNABERT, Nucleotide Transformer, and scGPT, we find that attention heads learn biologically meaningful associations during self-supervised pre-training and that these associations shift with fine-tuning. We show that interpretability varies with tokenization scheme, and that context-dependence plays a key role in head behaviour. Through ablation, we demonstrate that heads strongly associated with biological features are more important for task performance than uninformative heads in the same layers. In DNABERT trained for TATA promoter prediction, we observe heads with positive and negative associations reflecting positive and negative learning dynamics. Our results offer a framework to trace how biological features are learned from random initialization to pre-training to fine-tuning, enabling insight into how genomic foundation models represent nucleotides, genes, and cells.

bioinformatics↗

bioRxiv2024

Concerted modulation of spontaneous behavior and time-integrated whole-brain neuronal activity by serotonin receptors

Serotonin neurons from the raphe nuclei project across the entire brain and modulate diverse physiology and behavior by acting on over a dozen receptors. Here, we took a step towards dissecting this complex process by examining the effects of agonists and antagonists of four widely expressed serotonin receptors (2A, 2C, 1A, and 1B) on spontaneous mouse behavior, which we related to time-integrated whole-brain neuronal activity as assessed by the expression of Fos, a canonical immediate-early gene product. Low-dimensional representations of behavioral and Fos map data revealed the dominant factors of variation in each domain, captured predictable differences across drug groups, and enabled predictions of behavioral changes following perturbations in Fos maps and vice versa. Our study provides a rich resource describing the effects of manipulating serotonin receptors on animal behavior and whole-brain integrated neuronal activity. It also establishes an experimental and analysis paradigm for interrogating the relationship between behavior and neuronal activity across different time scales.

neuroscience↗

bioRxiv2021

Developmental Dynamics of RNA Translation in the Human Brain

The precise regulation of gene expression is fundamental to neurodevelopment, plasticity, and cognitive function. While several studies have deeply profiled mRNA dynamics in the developing human brain, there is a fundamental gap in our understanding of accompanying translational regulation. We perform ribosome profiling from more than 70 human prenatal and adult cortex samples across ontogeny and into adulthood, mapping translation events at nucleotide resolution. In addition to characterizing the translational regulation of annotated open reading frames (ORFs), we identify thousands of previously unknown translation events, including small open reading frames (sORFs) that give rise to human- and/or brain-specific microproteins, many of which we independently verify using size-selected proteomics. Ribosome profiling in stem cell-derived human neuronal cultures further corroborates these findings and shows that several neuronal activity-induced long non-coding RNAs (lncRNAs), including LINC00473, a primate-specific lncRNA implicated in depression, encode previously undescribed microproteins. Physicochemical analysis of these brain microproteinss identifies a large class harboring arginine-glycine-glycine (RGG) repeats as strong candidates for regulating RNA metabolism. Moreover, we find that, collectively, these previously unknown human brain sORFs are enriched for variants associated with schizophrenia. In addition to significantly expanding the translational landscape of the developing brain, this atlas will serve as a rich resource for the annotation and functional interrogation of thousands of previously unknown brain-specific protein products.

neuroscience↗

bioRxiv2021

Architect: a tool for producing high-quality metabolic models through improved enzyme annotation

MotivationConstraints-based modeling is a powerful framework for understanding growth of organisms. Results from such simulation experiments can be affected at least in part by the quality of the metabolic models used. Reconstructing a metabolic network manually can produce a high-quality metabolic model but is a time-consuming task. At the same time, current methods for automating the process typically transfer metabolic function based on sequence similarity, a process known to produce many false positives. ResultsWe created Architect, a pipeline for automatic metabolic model reconstruction from protein sequences. First, it performs enzyme annotation through an ensemble approach, whereby a likelihood score is computed for an EC prediction based on predictions from existing tools; for this step, our method shows both increased precision and recall compared to individual tools. Next, Architect uses these annotations to construct a high-quality metabolic network which is then gap-filled based on likelihood scores from the ensemble approach. The resulting metabolic model is output in SBML format, suitable for constraints-based analyses. Through comparisons of enzyme annotations and curated metabolic models, we demonstrate improved performance of Architect over other state-of-the-art tools. AvailabilityCode for Architect is available at https://github.com/ParkinsonLab/Architect. Contactjohn.parkinson@utoronto.ca Supplementary informationSupplementary data are available at Bioinformatics online.

systems biology↗

bioRxiv2020

A flexible repertoire of transcription factor binding sites and diversity threshold determines enhancer activity in embryonic stem cells

Transcriptional enhancers are critical for development, phenotype evolution and often mutated in disease contexts; however, even in well-studied cell types, the sequence code conferring enhancer activity remains unknown. We found genomic regions with conserved binding of multiple transcription factors in mouse and human embryonic stem cells (ESCs) contain on average 12.6 conserved transcription factor binding sites (TFBS). These TFBS are a diverse repertoire of 70 different sequences representing the binding sites of both known and novel ESC regulators. Remarkably, using a diverse set of TFBS from this repertoire was sufficient to construct short synthetic enhancers with activity comparable to native enhancers. Site directed mutagenesis of conserved TFBS in endogenous enhancers or TFBS deletion from synthetic sequences revealed a requirement for more than ten different TFBS. Furthermore, specific TFBS, including the OCT4:SOX2 co-motif, are dispensable, despite co-binding the OCT4, SOX2 and NANOG master regulators of pluripotency. These findings reveal a TFBS diversity threshold overrides the need for optimized regulatory grammar and individual TFBS that bind specific master regulators.

genomics↗

bioRxiv2019

A framework for spatial normalization and voxelwise analysis of diffusion propagators in multiple MAP-MRI data sets

We describe a pipeline for constructing a study-specific template of diffusion propagators measured with mean apparent propagator (MAP) MRI that supports direct voxelwise analysis of differences between propagators across multiple data sets. The pipeline leverages the fact that MAP-MRI is a generalization of diffusion tensor imaging (DTI) and combines simple and robust processing steps from existing tensor-based image registration methods. First, we compute a DTI study template which provides the reference frame and scaling parameters needed to construct a standardized set of MAP-MRI basis functions at each voxel in template space. Next, we transform each subjects diffusion data, including diffusion weighted images (DWIs) and gradient directions, from native to template space using the corresponding tensor-based deformation fields. Finally, we fit MAP coefficients in template space to the transformed DWIs of each subject using the standardized template of MAP basis functions. The consistency of MAP basis functions across all data sets in template space allows us to: 1. compute a template of propagators by directly averaging MAP coefficients and 2. quantify voxelwise differences between co-registered propagators using the angular dissimilarity, or a probability distance metric, such as the Jensen-Shannon Divergence. We illustrate the application of this method by generating a template of MAP propagators for a cohort of healthy volunteers and show a proof-of-principle example of how this pipeline may be used to detect subtle differences between propagators in a single-subject longitudinal clinical data set. The ability to standardize and analyze multiple clinical MAP-MRI data sets could improve assessments in cross-sectional and single-subject longitudinal clinical studies seeking to detect subtle microstructural changes, such as those occurring in mild traumatic brain injury (mTBI), or during the early stages of neurodegenerative diseases, or cancer.

neuroscience↗

bioRxiv2019

Coxiella burnetii small RNA 12 binds CsrA regulatory protein and transcripts for the CvpD type IV effector, regulates pyrimidine and methionine metabolism, and is necessary for optimal intracellular growth and vacuole formation during infection

Coxiella burnetii is an obligate intracellular gammaproteobacterium and zoonotic agent of Q fever. We previously identified 15 small non-coding RNAs (sRNAs) of C. burnetii. One of them, named CbsR12 (Coxiella burnetii small RNA 12) is highly expressed during growth in axenic medium and becomes even more dominant during infection of cultured mammalian cells. Secondary structure predictions of CbsR12 revealed four putative CsrA-binding sites in single-stranded segments of stem loops with consensus AGGA/ANGGA motifs. From this foundation, we determined that CbsR12 binds to recombinant C. burnetii CsrA-2, but not CsrA-1, proteins in vitro. Moreover, through a combination of in vitro and in vivo assays, we identified several in trans mRNA targets of CbsR12. Of these, we determined that CbsR12 binds to and upregulates translation of carA transcripts coding for carbamoyl phosphate synthetase A; an enzyme that catalyzes the first step of pyrimidine biosynthesis. In addition, CbsR12 binds and downregulates translation of metK transcripts coding for S-adenosyl methionine (SAM) synthase, a component of the methionine cycle. Furthermore, we found that CbsR12 binds to and downregulates the quantity of cvpD transcripts, coding for a type IVB effector protein, in vitro and in vivo. Finally, we found that CbsR12 is necessary for full expansion of Coxiella-containing vacuoles (CCVs) and affects bacterial growth rates in a dose-dependent manner in the early phase of infecting THP-1 cells. This is the first detailed characterization of a trans-acting sRNA of C. burnetii and the first example of a bacterial sRNA that regulates both CarA and MetK expression. CbsR12 is also one of only a few identified trans-acting sRNAs that interacts with CsrA. Results illustrate the importance of sRNA-mediated regulation in establishment of the intracellular CCV niche.\n\nAuthor summaryC. burnetii is an obligate intracellular bacterial pathogen that is transmitted to humans from animal reservoirs. Upon inhalation of aerosolized C. burnetii, the agent is phagocytosed by macrophages in the lung. The pathogen subverts macrophage-mediated degradation and resides in a large, intracellular, acidic vacuole, termed the Coxiella-containing vacuole (CCV). Small RNAs (sRNAs) are not translated into proteins. Instead, they target mRNAs in order to up- or down-regulate their stability and translation. Alternatively, some sRNAs bind to regulatory proteins and serve as \"sponges\" that effectively sequester the proteins and inhibit their function. C. burnetiis CbsR12 sRNA is highly expressed during infection in order to expand the CCV, and it works by a variety of mechanisms, including: 1) directly regulating transcripts of several metabolic genes that aid in bacterial replication, 2) binding to and regulating transcripts of a type IV effector protein that aids in infection, and 3) indirectly regulating an unknown number of genes by binding to a homolog of the global regulatory protein, CsrA. CbsR12 represents one of only a few sRNAs known to bind and sequester CsrA while also directly regulating mRNAs.

microbiology↗