2025-05-18 | | Total: 14
Spatial transcriptomics (ST) has shown remarkable promise in pathology applications, shedding light on the spatial organization of gene expression and its relationship to the tumor microenvironment. However, its clinical adoption remains limited due to the low throughput of current sequencing technologies. While recent methods attempt to infer ST from WSIs using pretrained image encoders, they remain constrained by limited gene coverage, organ-specific training, and the need for dataset-specific fine-tuning. In light of this, we introduce STPath, a generative foundation model pretrained on a large-scale collection of whole slide images (WSIs) with paired ST. Through extensive pretraining with tailored noise schedules, STPath learns representations across 38,984 genes and 17 organs, enabling accurate gene expression prediction without requiring further fine-tuning. STPath integrates different data modalities, including histological images, gene expressions, organ type, and sequencing technology information, within a geometry-aware Transformer architecture. We evaluate STPath across 5 tasks spanning 23 datasets and 9 biomarkers, including gene expression prediction, spot imputation, spatial clustering, gene mutation prediction, and survival prediction. The results demonstrate STPath’s strong ability to infer spatially resolved gene expression and reveal crucial pathological structures within tissue samples, underscoring its promise for scalable ST-based pathology applications.
Trehalose 6-phosphate (Tre6P) is a signal metabolite that links carbon metabolism with plant development. Our current understanding of Tre6P metabolism and signalling is predominantly based on studies in Arabidopsis thaliana. Some features could be adapted to the specific physiology, anatomy, and life cycle of this C3 eudicot model species, and thus might not be representative of other angiosperms. To better understand the regulation of carbon metabolism by Tre6P in monocot C4 species we used Setaria viridis, which has been widely adopted as a model for the major C4 NADP-malic enzyme subtype crop species, such as maize (Zea mays), sorghum (Sorghum bicolor) and sugarcane (Saccharum officinarum). In this work, we analysed the levels of transcripts encoding Tre6P-related enzymes in different tissues and cell types from S. viridis. The TREHALOSE-6-PHOSPHATE SYNTHASE1 transcript, its encoded protein (TPS1, the enzyme responsible for Tre6P synthesis) and Tre6P were mainly located in bundle sheath cells of S. viridis. Our results show that Tre6P is predominately synthesized and located in bundle sheath and associated cells, where it could play a fundamental role in the regulation of sucrose levels by modulating phloem loading. Highlight Trehalose 6-phosphate, a sugar signalling metabolite, is present mainly in the bundle sheath cells of Setaria viridis leaves, implicating it in the regulation of phloem loading.
MECP2 is commonly mutated in Rett syndrome, where MECP2’s function as a DNA cytosine methylation reader is believed critical. MECP2 variants are also catalogued in individuals with autism spectrum disorder (ASD), including nine missense variants with no known clinical significance. To assess these nine as risk alleles for ASD, we developed MECP2 variant function assays using yeast, Drosophila and human cell lines. We calibrated these assays with known reference pathogenic and benign variants. Our data predict that four ASD variants are loss of function (LoF) and five are functional. Protein destabilization or nuclear delocalization offers insight into the altered function of a number of these variants. Notably, yeast and Drosophila lack DNA methylation, yet all Rett reference pathogenic and ASD variants in the methyl DNA binding domain that we analyzed proved to be LoF, suggesting a clinically-relevant role for non-methyl DNA-binding by MECP2.
MeDIP-seq is an enrichment-based DNA methylation profiling technique that measures the abundance of methylated DNA. While this technique offers efficiency advantages over direct methylation profiling, it does not provide absolute quantification of DNA methylation necessary for cell type deconvolution. We introduce decemedip, a Bayesian hierarchical model for cell type deconvolution of methylated sequencing data that leverages reference atlases derived from direct methylation profiling. We demonstrate its accuracy and robustness through simulation studies and validation on cross-platform measurements, and highlight its utility in identifying tissue-specific and cancer-associated methylation signatures using MeDIP-seq profiling of patient-derived xenografts and cell-free DNA. decemedip is available at https://github.com/nshen7/decemedip.
Bacteria acquire new genes by horizontal gene transfer (HGT). Acquisition is typically mediated by mobile genetic elements (MGEs), however, beyond plasmids, bacteriophages and certain integrative conjugative elements (ICEs), the nature and diversity of MGEs is poorly understood. The bacterium Pseudomonas fluorescens SBW25 was propagated by serial transfer in the presence of filtrate obtained from garden compost communities. Genome sequencing of derived colonies revealed acquisition of three different mobile elements, each integrated immediately downstream of tmRNA. All are flanked by direct repeats and harbour a tyrosine integrase (intY), followed by a cargo of accessory genes including putative phage defence systems. Although characteristic of genomic islands, MGE-classifiers showed no matches to mobile elements. Interrogation of DNA sequence databases showed that similar elements are widespread in the genus Pseudomonas and beyond, with Vibrio Pathogenicity Island-1 (VPI-1) from V. cholerae being a notable example. Bioinformatic analyses demonstrate frequent transfer among diverse hosts. With focus on a single 55 kb element (I55) we show that intY is necessary for excision and circularisation, that the element is incapable of autonomous horizontal transfer, but is mobilizable – in the absence of direct cell-cell contact – upon addition of community filtrate. Further analyses demonstrate that I55 enhances host fitness in the presence of community filtrate, which stems in part from ability to defend against phages. Significance Statement The impact of horizontal gene transfer on the evolution of bacteria outpaces that driven by spontaneous mutation, but knowledge of the range of mediators, the genes mobilised, and the mechanisms of movement have largely depended on inferences stemming from bioinformatics. Here we describe a real-time evolution experiment in which a single focal strain propagated with filtrate from a complex microbial community captures genetic elements carrying a diverse cargo of genes whose mobility was previously unknown. The elements are representative of a class of mobile DNA that depend on nothing other than a tyrosine integrase that targets a highly conserved gene. The elements encode phage resistance, require the integrase to excise from the genome, confer a significant fitness advantage, and depend on a secondary element for transfer.
In eukaryotes, cellular respiration takes place in the cristae of mitochondria. The mitochondrial inner membrane protein Mic60, a core component of the mitochondrial contact site and cristae organizing system (MICOS), is crucial for the organization and stabilization of crista junctions and its associated functions. While the C-terminal Mitofilin domain of Mic60 is necessary for cellular respiration, the sequence determinants for this function have remained unclear. Here, we used ancestral sequence reconstruction to generate Mitofilin ancestors up to and including the last opisthokont common ancestor (LOCA). We found that yeast-lineage derived Mitofilin ancestors as far back as the LOCA rescue respiration. By comparing Mitofilin ancestors, we identified four residues sufficient to explain the respiratory difference between yeast- and animal-derived Mitofilin ancestors. Our results provide a foundation for investigating the conservation of Mic60-mediated cristae junction interactions.
This systematic and meta-analytical review examined how a reduction in oxygen availability to tissue (hypoxia) affects cognitive function. Hypoxia had a moderate-to-large detrimental effect on general cognitive ability and across domains, including memory, attention, executive function, processing speed, and psychomotor speed. Increased hypoxic severity was associated with greater declines in general cognitive ability and executive function, while longer duration of exposure was associated with greater declines in executive function and psychomotor speed. Participant age was a moderator for executive function and psychomotor speed, with older adults experiencing greater impairments. For executive function and psychomotor speed, the magnitude of these effects was less pronounced during intermittent and hypobaric exposures, potentially due to adaptive physiological mechanisms. While our models accounted for exposure characteristics and age of participants, substantial unexplained variance remained. These findings highlight hypoxias impact on cognition and emphasize the need to investigate underlying neurophysiological mechanisms that may influence individual vulnerability.
Methods utilizing k-mers are widely used in bioinformatics, yet our understanding of their statistical properties under realistic mutation models remains incomplete. Previously, substitution-only mutation models have been considered to derive precise expectations and variances for mutated k-mers and intervals of mutated and nonmutated sequences. In this work, we consider a mutation model that uses insertions and deletions in addition to single-nucleotide substitutions. Within this framework, we derive closed-form k-mer-based-estimators for the three fundamental mutation parameters: substitution rate, deletion rate, and average insertion length. We provide statistics of k-mers under this model and theoretical guarantees via concentration inequalities, ensuring correctness under reasonable conditions. Empirical evaluations on simulated evolution of genomic sequences confirm our theoretical findings, demonstrating that accounting for indel signals allows for accurate estimation of mutation rates and improves upon the results obtained by considering a substitution-only model. An implementation of estimating the mutation parameters from a pair of FASTA files is available here here: https://github.com/mahmudhera/estimate_rates_using_mutation_model.git The results presented in this manuscript can be reproduced using the code available here: https://github.com/mahmudhera/est_rates_experiments.git
1-Deoxysphingolipids are non-canonical sphingolipids linked to several diseases; however, their cellular effects are poorly understood. Here, we utilize lipid chemical biology approaches to investigate the role of 1-deoxysphingolipid metabolism on the properties and functions of secretory membranes. We first applied organelle-specific bioorthogonal labeling to visualize the subcellular distribution of metabolically tagged 1-deoxysphingolipids in RPE-1 cells, observing that they are retained in the endoplasmic reticulum (ER). We found that 1-deoxysphingolipids can be transported by the non-vesicular transporter CERT but are retained at ER exit sites (ERES), suggesting that they do not efficiently sort into vesicular carriers. Cells expressing disease-associated variants of serine palmitoyl-CoA transferase (SPT) accumulated 1-deoxysphingolipids, which reduced ER membrane fluidity and enlarged ERES. We observed that the rates of membrane protein release from the ER were altered in response to mutant SPT expression, dependent on the cargo’s affinity for ordered or disordered membranes. We propose that dysregulation of sphingolipid metabolism can alter secretory membrane properties, thereby affecting protein trafficking.
The interactions of nanomaterials with biomolecules in vivo determine their biological fate. Here, we show that a self-assembled peptide amphiphile nanostructure (namely SA-E) dynamically interacts with endogenous biomolecules and takes advantage of naturally occurring processes to target a broad range of solid tumors. Upon in vivo administration, self-assembled nanostructures of SA-E disassemble and reassemble with lipoproteins in circulation. Hitchhiking on lipoproteins prolongs the blood circulation of SA-E and allows it to cross endothelial barriers through transcytosis. At the tumor site, SA-E internalizes into cancer cells by mainly interacting with lipid-raft domains on cell membranes. By exploiting these endogenous interactions, SA-E demonstrated high tumor accumulation with extended retention in various xenograft, syngeneic, patient-derived xenograft, or transgenic mouse and rat models. In addition, SA-E enabled the effective delivery of highly potent chemotherapy to breast and glioma tumors with reduced side effects. With its simple and modular design and universal tumor accumulation mechanism, SA-E represents a promising platform for broad applications in cancer imaging and therapy.
Changing one’s mind is a complex cognitive phenomenon involving a continuous re-appraisal of the trade-off between past costs and future value. Recent work modeling this behavior across species has established associations between aspects of this choice process and their contributions to altered decision-making in psychopathology. Here, we investigated the actions in medial prefrontal cortex (mPFC) neurons of long intergenic non-coding RNA, LINC00473, known to induce stress resilience in a striking sex-dependent manner, but whose role in cognitive function is unknown. We characterized complex decision-making behavior in male and female mice longitudinally in our neuroeconomic foraging paradigm, Restaurant Row, following virus-mediated LINC00473 expression in mPFC neurons. On this task, mice foraged for their primary source of food among varying costs (delays) and subjective value (flavors) while on a limited time-budget during which decisions to accept and wait for rewards were separated into discrete stages of primary commitments and secondary re-evaluations. We discovered important differences in decision-making behavior between female and male mice. LINC00473 expression selectively influenced multiple features of re-evaluative choices, without affecting primary decisions, in female mice only. These behavioral effects included changing how mice (i) cached the value of the passage of time and (ii) weighed their history of economically disadvantageous choices. Both processes were uniquely linked to change-of-mind decisions and underlie the computational bases of distinct aspects of counterfactual thinking. These findings reveal a key bridge between a molecular driver of stress resilience and psychological mechanisms underlying sex-specific decision-making proclivities.
The advent of affordable whole-genome sequencing has spurred numerous large-scale projects aimed at inferring the tree of life, yet achieving a complete species-level phylogeny remains a distant goal due to significant costs and computational demands. Traditional species tree inference methods, though effective, are hampered by the need for high-coverage sequencing, high-quality genomic alignments, and extensive computational resources. To address these challenges, this study introduces WASTER, a novel de novo tool for inferring species trees directly from short-read sequences. WASTER employs a k-mer based approach for identifying variable sites, circumventing the need for genome assembly and alignment. Using simulations, we demonstrate that WASTER achieves accuracy comparable to that of traditional alignment-based methods, even for low sequencing depth, and has substantially higher accuracy than other alignment-free methods. We validate WASTER’s efficacy on real data, where it accurately reconstructs phylogenies of eukaryotic species with as low depth as 1.5X. WASTER provides a fast and efficient solution for phylogeny estimation in cases where genome assembly and/or alignment may bias analyses or is challenging, for example due to low sequencing depth. It also provides a method for generating guide trees for tree-based alignment algorithms. WASTER’s ability to accurately estimate trees from low-coverage sequencing data without relying on assembly and alignment will lead to substantially reduced sequencing and computational costs in phylogenomic projects.
GerS is a key lipoprotein regulator of Clostridioides difficile spore germination that shares structural, but not sequence, similarity to LolA in Gram-negative bacteria. GerS and LolA have vastly different biological roles: LolA functions in the trafficking of lipoproteins across the periplasmic space to the outer membrane of Gram-negative bacteria, while GerS allosterically activates the germination-specific amidase CwlD in endospore-forming Gram-positive bacteria. Here, we explore the diversity of LolA-like proteins across bacteria and reveal that GerS homodimer formation is a conserved feature of Peptostreptococcaceae family orthologs. Since dimerization of C. difficile GerS occurs via strand-swapping, we sought to investigate the functional importance of its strand swap by identifying mutations that alter the GerS dimer:monomer equilibrium. Our analyses reveal that GerSK71R and GerSN74L substitutions in the strand-swap hinge of GerS stabilize the homodimer, while a GerST72P substitution promotes monomer formation. Conversely, substituting the highly conserved proline at the equivalent site in LolA for threonine converts LolA from a monomer to a homodimer. Since our data indicate that destabilizing the GerS dimer impairs both GerS:CwlD binding in vitro and CwlD function in C. difficile, strand-swapped dimer formation in LolA-like proteins enhances their affinity for their binding partners. These data are consistent with analyses of equivalent LolA mutants in E. coli and thus provide new insight into the evolution and specialization of this intriguing group of proteins across the bacterial domain.
Asymmetric vertebrate heart development is driven by an intricate sequence of morphogenetic cell movements, the coordination of which requires precise interpretation of signaling cues by heart primordia. Here we show that Nodal functions cooperatively with FGF during heart tube formation and asymmetric placement. Both pathways act as migratory stimuli for cardiac progenitor cells (CPCs), but FGF is dispensable for directing heart tube asymmetry, which is governed by Nodal. We further find that Nodal controls CPC migration by inducing left-right asymmetries in the formation of actin-based protrusions in CPCs. Additionally, we define a developmental window in which FGF signals are required for proper heart looping and show cooperativity between FGF and Nodal in this process. We present evidence FGF may promote heart looping through addition of the secondary heart field. Finally, we demonstrate that loss of FGF signaling affects proper development of the atrioventricular canal (AVC), which likely contributes to abnormal chamber morphologies in FGF-deficient hearts. Together, our data shed insight into how the spatiotemporal dynamics of signaling cues regulate the cellular behaviors underlying organ morphogenesis. Summary statement This study explores the cooperative and independent roles of Nodal and FGF signaling in generating heart asymmetry.