Cell genomics最新文献

Generative genomic models hold immense potential for creating intricate biological systems, yet designing precise functional sequences remains challenging. Merchant et al. in Nature present semantic design, which employs the Evo genomic language model to generate novel functional genes based solely on genomic context. Moreover, SynGenome database houses over 120 billion sequences generated through semantic design, spanning a diverse range of functions.

We present a collection of 17 high-quality long-read inbred mouse strain genomes with complete annotation (contig N50s of 0.8-33.9 Mbp). This collection includes 12 widely used classical laboratory strains and 5 wild-derived strains. We have resolved previously incomplete genomic regions, including the major histocompatibility complex (MHC), defensin cluster, T cell receptor, and Ly49 complexes. Hundreds of non-reference genes from previous publications not found in GRCm39, such as Defa1, Raet1a, and Klra20 (Ly49T), were localized in the new reference genomes. We conducted a genome-wide scan of variable number tandem repeats (VNTRs) within the coding regions, identifying over 400 genes with VNTR polymorphisms with up to 600 repeat copies and repeat units reaching 990 nucleotides. Our strain-specific annotations enhance RNA sequencing (RNA-seq) analyses, as demonstrated in PWK/PhJ, where we observed a 5.1% improvement in read mapping and expression-level differences in 2.1% of coding genes compared to using GRCm39.

Recent algorithmic advances have enabled the inference of genome-wide ancestral recombination graphs (ARGs) from large genomic cohorts, providing detailed models of genealogical relatedness along the genome. These inferred ARGs can complement genotype imputation by capturing the effects of unobserved variants, but their use in large-scale linear mixed-model analyses has been computationally prohibitive. Here, we develop methods that leverage the ARG to perform genotype-matrix multiplications in sublinear time and implement scalable randomized algorithms for mixed-model analyses. We introduce ARG-RHE, a randomized Haseman-Elston approach for estimating narrow-sense heritability and performing region-based association testing using ARGs, enabling parallel analysis of multiple quantitative traits. Through extensive simulations, we demonstrate the computational efficiency and statistical power of this approach. Applied to 21,159 genes and 52 blood traits in 337,464 UK Biobank participants, ARG-RHE identifies 8% more gene-trait associations than imputation alone, demonstrating that genome-wide genealogies may be leveraged to complement genotype imputation in complex trait analyses.

Genetic mutations that yield a defective cystic fibrosis (CF) transmembrane regulator (CFTR) protein cause CF, a life-limiting autosomal-recessive Mendelian disorder. A protective role of CFTR loss-of-function mutations in inflammatory bowel disease (IBD) has been suggested, but its evidence has been inconclusive and contradictory. Here, leveraging a large IBD exome sequencing dataset comprising 38,558 cases and 66,945 controls of European ancestry in the discovery stage and a combined total of 42,475 cases and 192,050 controls across diverse ancestry groups in the replication stage, we established a protective role of CF-risk variants against IBD based on the association test of CFTR deltaF508 (p = 8.96E-11) and the gene-based burden test of CF-risk variants (p = 3.9E-07). Furthermore, we assessed variant prioritization methods, including AlphaMissense, using clinically annotated CF-risk variants as the gold standard. Our findings highlight the critical and unmet need for effective variant prioritization in gene-based burden tests.

Cellular senescence, a stress-induced program causing stable cell-cycle arrest, is a hallmark of liver aging, fibrosis, and cancer. However, the cell-type-specific mechanisms, spatial organization, and cancer-associated alterations in the liver remain unclear. We profiled 43 normal human livers spanning ages and fibrosis stages using a single-cell multiome, Xenium spatial transcriptomics, and CODEX, complemented by fibrotic mouse models and 24 colorectal cancer liver metastases. We found CDKN1A+ senescent hepatocytes, fibroblasts, cholangiocytes, and endothelial cells associated with age, liver disease, or cancer. Senescence differed between aged and fibrotic livers, with similar patterns in mice. Spatially, CDKN1A+ hepatocytes localized periportally, while SERPINE1+ aging-associated hepatocytes formed spatial clusters, potentially mediated by Claudins and THBS1. Fibrotic regions contained CXCL12+ senescent fibroblasts interacting with CXCR4+ immune cells. Chemotherapy intensified senescence in hepatocytes by 5-fold relative to aging and led to unique CDKN2A+ populations. Across conditions, senescent cells shared AP-1 activation, pro-inflammatory cytokines, and apoptosis resistance, suggesting therapeutic opportunities.

Phenotypic variation arises from interactions between genetic and environmental factors, but disentangling these effects for complex traits remains challenging in observational cohorts like human biobanks. Model organisms with controlled genetic and environmental variation complement human studies in analyzing higher-order effects such as gene-by-environment (G×E) interactions, dominance, and epistasis. We utilized 76 medaka strains from the Medaka Inbred Kiyosu-Karlsruhe (MIKK) panel to compare heart rate plasticity across temperatures. An F2 segregation analysis identified 16 quantitative trait loci (QTLs), many exhibiting dominance, G×E, G×G, and G×G×E interactions. We experimentally validated four candidate genes, revealing temperature-sensitive heart rate effects. Finally, we simulated how genome-wide association study (GWAS) discovery power depends on statistical model choice. Our results suggest that the limited detection of non-additive effects in human GWASs stems from current study designs and sample sizes. This work demonstrates the value of controlled model organism studies for dissecting complex trait genetics and informing association study design.

The nematode Caenorhabditis elegans is a favorable model for studying cell-surface protein interactomes, given its well-defined and stereotyped intercellular contacts. Here, we report an extracellular interactome dataset for C. elegans. Most of these interactions were unknown, despite recent datasets for flies and humans, as our collection contains a larger selection of protein families. We uncover interactions for all four major axon guidance pathways, including ectodomain interactions between three of them. We demonstrate that a protein family, previously known for maintaining axon positioning, functions as secreted binders for insulins and that their overexpression in vivo extends lifespan, consistent with inhibition of insulin signaling. We reveal interactions of cystine-knot proteins with putative signaling receptors, which may extend the study of neurotrophins and growth factors to nematodes. Finally, our dataset constitutes a resource for uncovering the logic of neuronal connectivity, intercellular communication and adhesion, and signaling pathways involved in aging and disease.

Recent breakthroughs in spatial multi-omics enable simultaneous profiling of different modalities while preserving tissue architecture, providing unprecedented opportunities to explore tissue complexity. However, due to the sparse and noisy nature of the data, interpreting these complex tissue structures and cellular communication remains challenging. We present MultiSP, a deep learning framework that enhances data representation through efficient spatial and feature similarity fusion, modality-specific probabilistic generative modeling, and cross-modality adversarial learning. Applied to various spatial multi-omics datasets, it outperforms existing methods in capturing biologically interpretable spatial domains. MultiSP also denoises spatial data, uncovers modality-specific spatial variations, and reveals gene regulation mechanisms. In the tumor microenvironment, it unravels fine-resolution cellular distribution maps, such as spatially neighboring macrophage-enriched sub-regions with distinct prognosis outcomes. Additionally, MultiSP facilitates the inference of spatially multimodal cell-cell communication. Together, MultiSP serves as a powerful framework for uncovering spatially multimodal heterogeneity and communication by integrating complementary information from multiple modalities.

Many cancers upregulate the expression of sialic acid-containing glycans. These oligosaccharides engage inhibitory sialic acid-binding immunoglobulin-like lectin (Siglec) receptors on immune cells, allowing cancer cells to evade immune surveillance. The genetic mechanisms underlying this process remain poorly defined. In this study, we performed gain-of-function CRISPR activation (CRISPRa) screens to define genetic pathways that regulate expression of Siglec-binding glycans. We show that Siglec ligand expression is controlled through genetic competition between genes that catalyze α2-3 sialylation and GlcNAcylation of galactose residues. Cancer glycome remodeling is also aided by the overexpression of "professional ligands" that facilitate Siglec-glycan binding. Notably, we also find that expression of the CD24 gene is genetically dispensable for cell surface binding of the inhibitory receptor Siglec-10. Finally, we identify the sulfotransferase enzyme GAL3ST4 as a potential driver of immune evasion in glioma cells. Our study provides a unique genomic atlas of cancer-associated glycosylation and identifies immediately actionable targets for cancer immunotherapy.

Gene expression is shaped by transcriptional regulatory networks (TRNs), where transcription regulators interact within regulatory elements in a context-specific manner. Deciphering context-specific TRNs has long been constrained by the severe sparsity of cell-type-specific chromatin immunoprecipitation sequencing (ChIP-seq) profiles. Here, we present ChromBERT, a foundation model pre-trained on large-scale human ChIP-seq datasets covering ∼1,000 transcription regulators. ChromBERT learns the genome-wide syntax of regulatory cooperation and generates interpretable TRN representations. After prompt-enhanced fine-tuning, it outperforms existing methods for imputing unseen cistromes. Moreover, lightweight fine-tuning on cell-type-specific downstream tasks adapts the TRN representations to capture regulatory effects and dynamics within any given cellular context. The resulting context-specific representations can then be interpreted to infer regulatory roles of transcription regulators underlying these cell-type-specific regulatory outcomes without requiring additional ChIP-seq experiments. By overcoming the limitations of sparse transcription regulator data, ChromBERT significantly enhances our ability to model and interpret transcriptional regulation across a wide range of biological contexts.