Background Upper tract urothelial carcinoma (UTUC) is an uncommon but aggressive malignancy. Muscle invasion is strongly associated with poor prognosis and limited treatment options. Revealing the molecular basis of muscle invasiveness in UTUC is crucial. Methods We characterised somatic structural variants (SSVs) from 162 patients with UTUC and integrated these data with bulk RNA-seq, single cell RNA-seq and spatial transcriptomics, with selected SSVs validated by single-molecule sequencing. Furthermore, functional validation experiments, including dual-luciferase reporter assays, wound healing, and invasion/migration assays, were conducted on both the SSV region upstream of TPX2 and the TPX2 gene itself in the 5637 cell line. Findings We found that SSVs were enriched in muscle-invasive (MI) -UTUC compared with non-muscle-invasive (NMI)-UTUC, and were associated with gene expression changes independent of copy-number variation. TPX2 emerged as a representative target, with TPX2-associated SSVs linked to TPX2 over expression and patient poorer progression-free and overall survival. Functional study indicated that a TPX2-upstream SSV region enhances TPX2 expression through promoter regulation, leading to increased invasiveness of 5637 cells. Single cell analyses revealed that TPX2-positive cycling epithelial cells exhibited heightened proliferative signalling and distinct ligand-receptor interactions, particularly involving EGFR- and EPHA2-related pathways. Spatial transcriptomics further localised TPX2-positive spots to epithelial mesenchymal transition (EMT) -enriched neighbourhoods with elevated EGFR/EPHA2 associated ligands, and showed markedly higher abundance in MI-UTUC. Interpretation SSVs drive transcriptional reprogramming, disease aggressiveness and microenvironmental remodelling in UTUC. The signatures of SSVs could classify UTUC into classes with different prognosis. Furthermore, SSVs of TPX2 might be a potential biomarker and candidate therapeutic vulnerability. Funding Stated in acknowledgements section of manuscript.

The deep sea, as the largest and maybe most hostile environment on Earth, is still underexplored, especially regarding its genetic repertoire. Yet, previous work has revealed significant habitat-specific deep-sea biodiversity. Here, we present an integrated deep-sea microbial genetic dataset comprising 502 million nonredundant genes from 2,138 samples and 2.4 million predicted structures and use it to link specific protein structures with genetic variants associated with life in the deep sea and to assess their biotechnology potential. Combining global sequence analysis with biophysical and biochemical measurements revealed unprecedented sequence diversity and substantial structural conservation of proteins. Especially, proteins involved in replication, recombination, and repair were identified as being under rapid evolution and with specialized properties. Among these, a structurally divergent helicase exhibited advantages in controlling nanopore sequencing speed. Thus, our work positions the deep sea as an evolutionary engine that generates and hosts genetic diversity and bridges genetic knowledge with biotechnology.

Land plants underpin civilization and planetary health, yet their genomic diversity remains largely uncharted. Current resources are unstandardized and scarce, lacking reference genomes for 95% of genera, 70% of families, and 51% of orders, impeding evolutionary and functional insight. We thus propose the PLANeT initiative, an international effort to generate high-quality, standardized genomes across the plant tree of life. Integrating artificial intelligence (AI) with genomics, we will decode conserved principles to advance fundamental plant biology, biodiversity conservation, crop improvement, and natural product discovery. Engaging around 100 labs to train 1,000 scientists, we will tackle pivotal questions for a sustainable future.

3020 Background: The drug failure rate has increased to ~95%, despite the growth in targeted therapies. As clinical trials demonstrated, a targeted gene alone does not predict whether patients have longer life expectancy in response to the drug. As studies with model organisms showed, the effect of the drug, and the mechanisms underlying it, depend on the entire multi-ome. But multi-omic data are small-cohort, noisy, and high-dimensional, i.e., extremely difficult to model. Methods: We have developed our artificial intelligence and machine learning (AI/ML) to overcome these challenges [doi: 10.1073/pnas.0530258100, 10.1158/1538-7445.AM2025-CT227]. We demonstrated our algorithms in the unsupervised modeling of, e.g., whole genomes of 85 astrocytoma patients. Mechanistic interpretation showed that the modeling blindly removed batch effects, separated normal demographic variations, and discovered a disease-specific genome-wide pattern of DNA copy-number alterations. This pattern was used to derive an actionable predictor of patients’ overall survival (OS) and gene targets to sensitize their tumors. We computationally validated both the predictor and the modeling in federated studies of mutually-exclusive sets of 59–251 patients. The modeling repeatedly discovered a representation of the predictor in every study, across astrocytoma grades II, III, and IV, i.e., glioblastoma (GBM), patients. We experimentally validated the predictor in a clinical trial of 79 GBM patients, initially retrospectively, and, in a four-year follow up, also prospectively [doi: 10.1063/1.5142559, 10.1145/3624062.3624078, 10.1200/JCO.2024.42.16_suppl.e14028]. In all the cohorts, the predictor, with 75–95% concordance with OS, was more accurate than all standard-of-care indicators. With 100% reproducibility among Complete Genomics, Illumina, and Ultima whole-genome sequencing, and > 99% when including Affymetrix and Agilent DNA microarrays, the predictor was also the most precise. Results: Here, we describe functional genomic experimental validation of both a predicted gene target and the predicted tumors’ responses to the targeting. Guide RNAs were designed and a lentiviral CRISPR-Cas9 all-in-one vector was utilized to knock out the modeling-predicted target METTL2A . Knockout validation at the protein level was performed using Western blot. Knockout in the patient-derived GBM cell lines U-87 MG and U-118 MG resulted in significantly attenuated cell viability and proliferation. The level of attenuation was significantly different between the cell lines, consistent with their whole genome-based predicted responses. Conclusions: Our quantum mechanics-based multi-tensor AI/ML solved the 75-year-old problem of correctly predicting — patients’ OS, drug responses, and gene targets — from their GBM tumors' whole genomes.

Animal hearts display diverse anatomical structures during adaptive evolution. Here, we present a multiomics atlas of adult hearts from 27 species across chordates, arthropods, and mollusks. Joint analysis indicates that Bilateria hearts share a core gene repertoire, taking a stepwise “add-on” approach as a universal evolutionary strategy. The “proto-heart” is populated by key cell types, including cardiomyocytes, fibroblasts, endothelial cells, and neural cells, which maintained core signatures while evolving with shifts in living environments and corresponding adaptations in the cardiovascular system. Additionally, we reveal an evolutionarily conserved cardiomyocyte state dynamic potentially linked to cardiac development and stress responses. Finally, we identify a common molecular program underpinning chamber evolution from a ventricular foundation. This work establishes a resource for understanding the intrinsic mechanisms of heart evolution.

The cellular innovations underlying vertebrate gastrointestinal diversification remain largely unknown. Here, we constructed a single-cell and spatial transcriptomic atlas comprising nearly 1 million cells from 10 species, spanning 500 million years of vertebrate evolution. Analysis of conserved gene co-expression modules shows that the repurposing of existing genetic programs, together with de novo gene emergence, drives cellular evolution. We link module remodeling to lineage-specific dynamics, including ciliated cell loss and the tuft cell emergence. Furthermore, we identify lymphoid aggregates in lungfish, suggesting that intestinal immune organization may be prior to the tetrapod lineage, and demonstrate that fish oxynticopeptic cells exhibit functional polarization that prefigures mammalian gastric specialization. Our atlas provides a global view of gastrointestinal evolution, highlighting the role of regulatory repurposing in defining organ function.

Abstract Adenosine-to-inosine (A-to-I) RNA editing is a widespread post-transcriptional mechanism that diversifies the transcriptome. While ADAR enzymes catalyze this reaction, the upstream mechanisms that determine why individual adenosines are edited at markedly different efficiencies remain poorly understood. Here, we integrated multi-omics datasets from human cell lines and mouse embryonic tissues and developed machine learning models that distinguish high-and low-frequency editing sites based on local epigenetic features. Across species, tissues, and developmental stages, H3K36me3 consistently emerges as the strongest negative predictor of editing frequency, whereas the histone variant H2A.Z.1 shows a positive association. Functional validations in H2AFZ-knockdown cells reveal that H2A.Z.1 preferentially facilitates editing at high-frequency sites (76.3% of sites decreased, mean Δ =-0.07), whereas SETD2-knockout-mediated loss of H3K36me3 selectively derepresses editing at low-frequency sites (86.9% upregulated, ∼2.7-fold increase). These findings establish chromatin state as an upstream regulatory layer that modulates RNA editing independently of editing enzyme abundance and reveal opposing roles for H3K36me3 and H2A.Z.1 in shaping RNA editing landscapes. Together, our study provides a conceptual framework linking epigenetic regulation to post-transcriptional RNA modification, suggesting that chromatin-mediated regulation contributes to the establishment of site-specific RNA editing programs across mammalian genomes.

The vaginal microbiome is essential for women's health, yet its genomic diversity and interaction with the host remain incompletely characterized. Here we present the Global Vaginal Metagenome-assembled Genomes catalog, an extensive repository of vaginal microbial genomes generated by integrating 10,665 in-house Chinese metagenomes, with 2,967 publicly available metagenomes and 1,433 bacterial isolates. The catalog comprises 65,055 genomes from 890 prokaryotes, 11 eukaryotes and 6,590 viral taxonomic units, many not represented in public reference databases. We investigate virus-bacteria interactions, revealing conserved phages-host associations. We then identify substantial intraspecies genomic and functional variations displaying population-specific patterns. A metagenome-genome-wide association study identifies seven host genetic loci associated with vaginal species at study-wide significance and replicated in at least one independent cohort, notably connecting the gene OPRK1 with the potential pathogen Ureaplasma urealyticum. In summary, our research provides a comprehensive reference for future studies on genotype-phenotype interplay within the human vaginal microbiome.

Regeneration relies on precise spatiotemporal gene expression and cellular responses to establish tissue identity and body patterning. Using high-resolution Stereo-seq (715 nm) on 353 sections from 16 whole animals at 8 regeneration timepoints, we constructed a 4D spatiotemporal transcriptomic map of planarian regeneration. Our analysis captured 36 refined cell types from 3,508,004 segmented cells, enabling genome-wide transcriptional imputation of gene expression dynamics across body axes at cellular, tissue, and organismal scales. We identified dynamic positional gradients and distinct spatially distributed cell types during regeneration, including an injury-induced Anterior Regenerative Zone (ARZ). The ARZ exhibited enriched positional signals in epidermal, muscle, and neural cells and was regulated by Mediator 8, which is crucial for polarity remodeling and blastema formation. This study provides a comprehensive spatial molecular and cellular map of regenerative processes, highlighting injury-induced spatial domains and key regulatory factors in planarian regeneration. We also provide an interactive web portal, offering a valuable resource for exploring and analyzing regeneration mechanisms in a spatiotemporal context.

Background Tumor evolution is driven by substantial cellular heterogeneity, yet its reconstruction remains challenging with conventional bulk sequencing approaches. Single-cell transcriptomic data provide opportunities to study clonal diversity and evolutionary dynamics, but accurate inference of clonal structure and lineage relationships from these data remains difficult. Methods We developed single-cell reinforcement learning (RL) for evolution modeling (scRevol), an RL-based model for inferring tumor evolution from single-cell RNA sequencing (scRNA-seq) data. Using copy number variation (CNV) profiles inferred from scRNA-seq data, scRevol employs a label assignment learning strategy to generate informative embeddings, identify clonal populations, and reconstruct evolutionary trajectories. We evaluated scRevol using simulated datasets, lineage tracing data, and ovarian cancer scRNA-seq datasets. Results In simulated datasets, scRevol showed robust intra-cluster coherence and accurately recovered lineage topology across varying levels of clonal complexity and noise. Compared with clustering baselines and existing methods for single-cell tumor evolution analysis, tscRevol achieved strong agreement with ground truth. In lineage tracing data, scRevol identified clonal groups associated with metastatic potential and revealed substantial metastatic heterogeneity. In ovarian cancer datasets, scRevol resolved subclonal structures across primary and metastatic lesions and associated inferred clones with distinct transcriptional and pathway-level features. Conclusions These results support scRevol as a practical framework for reconstructing tumor evolution from single-cell transcriptomic data and for characterizing clonal architecture and subclonal diversity.

BACKGROUND:Understanding how organisms reconstruct complex tissue architectures following injury requires precise mapping of gene expression and cellular responses across space and time. Although planarians serve as a classic model for whole-body regeneration, capturing the continuous spatiotemporal dynamics of positional information and cell fate decisions at the organismal scale remains a significant challenge. RESULTS:Using high-definition spatial transcriptomics, we generated a 4-dimensional atlas encompassing over 3.5 million cells from whole animals across 8 distinct regeneration timepoints. This comprehensive dataset enabled the definition of 36 spatial domains and the tracing of body axis restoration, revealing that positional control genes recover through self-organizing dynamics analogous to an underdamped control system. We identified an injury-induced spatial domain termed the anterior regenerative zone. This unique region is characterized by the convergence of epidermal, muscular, and neural lineages enriched with positional signals. Furthermore, we demonstrated that the transcriptional co-factor Mediator 8 is a critical regulator of this zone. Depletion of Mediator 8 impairs the formation of the anterior regenerative zone, disrupts polarity establishment, and prevents successful blastema formation. CONCLUSIONS:Our study provides a holistic molecular and cellular reconstruction of whole-body regeneration, directly linking dynamic gene expression gradients to morphological restoration. The discovery of the Mediator 8-regulated anterior regenerative zone highlights the importance of spatial domains in coordinating tissue repair. The resulting interactive atlas serves as a foundational resource for deciphering the logic of spatiotemporal patterning in regeneration.

UNC93B1 is a crucial chaperone protein for the trafficking of Toll-like receptors (TLRs) and regulates antigen presentation in dendritic cells (DCs), which activates downstream immune responses. Here, we identified a novel homozygous gain-of-function (GOF) UNC93B1 variant in an early-onset lupus patient. The patient presented with an elevated level of inflammation and auto-antibodies, and organ damage. The Unc93b1R95L/R95L transgenic mice also exhibited autoimmune and autoinflammatory phenotypes. The transcriptional analysis revealed increased inflammation and elevated activation of DCs in the patient's peripheral blood mononuclear cells and bone marrow-derived DCs from Unc93b1R95L/R95L mice. In addition to the selected TLR7/8 activation in previously reported UNC93B1 GOF variants, the single-cell transcriptome and flow cytometry of splenocytes from Unc93b1R95L/R95L mice demonstrated increased phagocytosis activity and T helper cell differentiation with altered ICAM and MHC signaling in DCs and T cells, respectively. These results suggest that the UNC93B1 GOF variant enhances antigen presentation from DCs to T cells in the pathogenesis of immune dysregulation. Our study expands the pathogenic variant spectrum of UNC93B1 and offers insight into the underlying mechanism of antigen presentation in immune dysregulation caused by UNC93B1 beyond TLR trafficking dysfunction.

The spatial organization of cell types and gene expression underlies tissue architecture and organismal physiology. However, resolving complete cellular and molecular landscapes within a three-dimensional context remains challenging, particularly in organisms with high cellular heterogeneity and complex geometry. Here, we developed a semi-supervised workflow to reconstruct high-resolution three-dimensional spatial transcriptomic models of the planarian Schmidtea mediterranea at single-cell resolution. Planarians are basal bilaterians capable of regenerating body structures after injury. Our reconstruction maps the distribution of cell types and spatially patterned gene expression across the intact organism and identifies 119 candidate positional control genes (PCGs). These genes are expressed across muscle, neural, and epidermal populations. Functional perturbation of selected candidates, including pitx3 , ptpn11 , pi4ka , upf3b , and cul1 , revealed roles in regeneration following injury. In addition, we identify intestinal cells as prominent components of the microenvironment surrounding adult tgs1 ⁺ pluripotent stem cells (neoblasts). Together, these results demonstrate the utility of three-dimensional spatial transcriptomic reconstruction for mapping cellular architecture and positional gene regulation in complex adult organisms.

Background Cystic echinococcosis, caused by the tapeworm Echinococcus granulosus sensu stricto (ss), is a globally distributed, zoonotic disease that is recognised by WHO as a neglected tropical disease. Despite its clinical and economic importance, nuclear genomic variation in this parasite has not been systematically characterised across global populations. In this study, we aimed to characterise the genome-wide nuclear genetic diversity and population structure of E granulosus ss across globally distributed populations. Methods We conducted a genomic study of 137 E granulosus ss samples from endemic regions across five continents, derived from previously collected parasite material from livestock, wildlife, and human infections. Using a chromosome-scale reference genome, we applied population genomic approaches to investigate genome-wide nuclear genetic diversity, population structure, and patterns of evolutionary constraint. Findings We identified 1 071 085 nuclear single-nucleotide polymorphisms across 137 samples, with heterozygosity ranging from 46% to 93% per sample. Genome-wide analyses identified two major clades associated with geographical origin. Distinct regions of genetic differentiation were observed, particularly on chromosome 9. Conserved genes under purifying selection included those involved in glycan biosynthesis and core cellular functions, whereas variable genes were enriched in pathways such as ribosome biogenesis. Mitochondrial genotypes (G1 and G3) did not align with the nuclear genomic structure. Interpretation To the best of our knowledge, this study provides the first broad atlas of nuclear genomic diversity in E granulosus ss, uncovering genetic diversity and population structure. The findings have important implications for molecular epidemiology, genomic surveillance, and translational development of diagnostics and vaccines. Incorporating genomic data into cystic echinococcosis control programmes could enhance WHO-aligned efforts to reduce the burden of this neglected tropical disease. Funding Australian Research Council and the Estonian Ministry of Education and Research.

Abstract Despite the growth in targeted therapy development, the drug failure rate has increased to ∼95%. As clinical trials demonstrated, the targeted gene alone does not predict whether patients would have longer life expectancy in response to a drug. As studies with model organisms showed, the effect of the drug, and the mechanisms underlying it, depend on the entire multi-ome. But multi-omic data are small-cohort, noisy, and high-dimensional, i.e., extremely difficult to model. We have developed our quantum mechanics-based artificial intelligence and machine learning (AI/ML) to overcome these challenges [doi: 10.1158/1538-7445.AM2025-CT227]. We demonstrated our algorithms in the unsupervised modeling of, e.g., whole genomes of 85 astrocytoma patients. Mechanistic interpretation showed that the model blindly removed batch effects, separated normal demographic variations, and discovered a disease-specific genome-wide pattern of DNA copy-number alterations. This pattern was used to derive an actionable predictor of patients’ overall survival (OS) and gene targets to sensitize their tumors. We computationally validated both the predictor and the modeling in federated studies of mutually-exclusive sets of ∼50-250 patients. The modeling repeatedly discovered a representation of the predictor in every study, in astrocytoma grades II, III, and IV, i.e., glioblastoma (GBM), patients. We experimentally validated the predictor in a clinical trial in 79 GBM patients, initially retrospectively, and, in a four-year follow up, also prospectively [doi: 10.1063/1.5142559, 10.1145/3624062.3624078]. In all the cohorts, the predictor, with 75-95% concordance with survival, was more accurate than all standard-of-care indicators. With 100% reproducibility among Complete Genomics, Illumina, and Ultima whole-genome sequencing, and >99% when including Affymetrix and Agilent DNA microarrays, the predictor was also the most precise. Here, we describe functional genomics experimental validation of both a gene target predicted to sensitize the tumors, as well as the predicted tumors’ response level. Guide RNAs were designed and a lentiviral CRISPR-Cas9 all-in-one vector was utilized to knock out the candidate target. Knockout validation at the protein level was performed using Western blot. Knockout in patient-derived GBM cell lines resulted in significantly attenuated cell viability and proliferation. The level of attenuation was significantly different between the cell lines, in agreement with their whole genome-based predicted response. We conclude that our quantum mechanics-based multi-tensor AI/ML solved the 75-year-old problem of correctly predicting — GBM patients’ OS, gene targets to sensitize the tumors, and the tumors’ response to their targeting — from their whole genomes. Citation Format: Orly Alter, Sri Priya Ponnapalli, Marissa Coppola, Angela C. Gushue, Tessa O. House, Penelope L. Miron, Kristy L. Miskimen, Kristin A. Waite, Sarah Pollock, David Bogumil, Nika Iremadze, Samantha Hernandez, Nadiya Sosonkina, Sara E. Coppens, Anthony C. Bryan, Estevan P. Kiernan, Huanming Yang, Jay Bowen, Ghunwa A. Nakouzi, Doron Lipson, Jill S. Barnholtz-Sloan, Andrew E. Sloan, Tiffany R. Hodges, Asaf Zviran, Jessica W. Tsai. Quantum mechanics-based multi-tensor AI/ML correctly predicts — glioblastoma patients’ overall survival, gene targets to sensitize the tumors, and the tumors’ response to their targeting — from their whole genomes [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6884.

A comprehensive atlas of genes, cell types, and their spatial distribution across a whole mammalian brain is fundamental for understanding the function of the brain. Here, using single-nucleus RNA sequencing (snRNA-seq) and Stereo-seq techniques, we generated a mouse brain atlas with spatial information for 308 cell clusters at single-cell resolution, involving over 4 million cells, as well as for 29,655 genes. We have identified cell clusters exhibiting preference for cortical subregions and explored their associations with brain-related diseases. Additionally, we pinpointed 155 genes with distinct regional expression patterns within the brainstem and unveiled 513 long non-coding RNAs showing region-enriched expression in the adult brain. Parcellation of brain regions based on spatial transcriptomic information revealed fine structure for several brain areas. Furthermore, we have uncovered 411 transcription factor regulons showing distinct spatiotemporal dynamics during neurodevelopment. Thus, we have constructed a single-cell-resolution spatial transcriptomic atlas of the mouse brain with genome-wide coverage.

BACKGROUND:Prior studies on toxicants and CVD focused on limited compounds, restricting their relevance to real-world exposure scenarios, a comprehensive assessment multiple toxicants is therefore warranted. METHODS:An exposome-wide association study (ExWAS) was conducted using 61 toxicants in 10 categories among 3577 participants from the 2013-2016 NHANES. Weighted quantile sum (WQS) regression and Bayesian kernel machine regression (BKMR) were applied to identify dominant contributors. Mediation analysis and network toxicology were used to explore the mediating roles of the systemic inflammatory response index (SIRI) and phenotypic age acceleration. RESULTS:ExWAS identified 29 toxicants from eight chemical families positively associated with CVD risk. WQS regression revealed that mixed exposures to toxicant family of nicotine metabolites (aOR = 1.45, 95 % CI: 1.18-1.78), metals (aOR = 2.16, 95 % CI: 1.35-3.45), polycyclic aromatic hydrocarbons (PAHs) (aOR = 1.34, 95 % CI: 1.07-1.69), perchlorate/nitrate/thiocyanate (aOR = 1.26, 95 % CI: 1.03-1.53), and volatile organic compound metabolites (VOCs) (aOR = 1.33, 95 % CI: 1.03-1.73) were significantly associated with CVD risk, and BKMR results indicated that metals showed the highest posterior inclusion probability (1.000), followed by PAHs (0.655) and VOCs (0.591). Mediation analysis revealed that SIRI and phenotypic age acceleration significantly mediated the associations of 21 and 18 toxicants with CVD, respectively. Network toxicology analysis demonstrated that toxicants were enriched in multiple inflammaging-related pathways. CONCLUSION:When multiple toxicants are considered simultaneously, the metals, PAHs, and VOCs families contribute most substantially to CVD risk, and inflammaging emerged as mediator of the associations.

Animal behavior is linked to the gene regulatory network (GRN) coordinating gene expression in the brain. Eusocial honeybees, with their natural behavioral plasticity, provide an excellent model for exploring the connection between brain activity and behavior. Using single-nucleus RNA sequencing and spatial transcriptomics, we analyze the expression patterns of brain cells associated with the behavioral maturation from nursing to foraging. Integrating spatial and cellular data uncovered cell-type and spatial heterogeneity in GRN organization. Interestingly, the stripe regulon is explicitly activated in foragers' small Keyon cells, which are implicated in spatial learning and navigation. When worker age is controlled in artificial colonies, stripe and its key targets remained highly expressed in the KC regions of bees performing foraging tasks. These findings suggest that specific GRNs coordinate individual brain cell activity during behavioral transitions, shedding light on GRN-driven brain heterogeneity and its role in the division of labor of social life.

Over 320 million years of evolution, amniotes have developed complex brains and cognition through largely unexplored genetic and gene expression mechanisms. We created a comprehensive single-cell atlas of over 1.3 million cells from the telencephalon and cerebellum of turtles, zebra finches, pigeons, mice, and macaques, employing single-cell resolution spatial transcriptomics to validate gene expression patterns across species. Our study identifies significant species-specific variations in cell types, highlighting their conservation and diversification in evolution. We found pronounced differences in telencephalon excitatory neurons (EXs) and cerebellar cell types between birds and mammals. Birds predominantly express SLC17A6 in EX, whereas mammals express SLC17A7 in the neocortex and SLC17A6 elsewhere, possibly due to loss of function of SLC17A7 in birds. Additionally, we identified a bird-specific Purkinje cell subtype (SVIL+), implicating the lysine-specific demethylase 11 (LSD1)/KDM1A pathway in learning and circadian rhythms and containing numerous positively selected genes, which suggests an evolutionary optimization of cerebellar functions for ecological and behavioral adaptation. Our findings elucidate the complex interplay between genetic evolution and environmental adaptation, underscoring the role of genetic diversification in the development of specialized cell types across amniotes.

With rapid advancements in single-cell RNA sequencing (scRNA-seq) technologies, exploration of the systemic coordination of critical physiological processes has entered a new era. Here, we generated a comprehensive Arabidopsis single-nucleus transcriptomic atlas using over 1 million nuclei from 20 tissues encompassing multiple developmental stages. Our analyses identified cell types that have not been characterized in previous single-protoplast studies and revealed cell-type conservation and specificity across different organs. Through time-resolved sampling, we revealed highly coordinated onset and progression of senescence among the major leaf cell types. We originally formulated two molecular indexes to quantify the aging state of leaf cells at single-cell resolution. Additionally, facilitated by weighted gene co-expression network analysis, we identified hundreds of promising hub genes that may integratively regulate leaf senescence. Inspired by the functional validation of identified hub genes, we built a systemic scenario of carbon and nitrogen allocation among different cell types from source leaves to sink organs.