Search

Abstract Background Inhomogeneous patterns of chromatin-chromatin contacts within 10–100-kb-sized regions of the genome are a generic feature of chromatin spatial organization. These features, termed topologically associating domains (TADs), have led to the loop extrusion factor (LEF) model. Currently, our ability to model TADs relies on the observation that in vertebrates TAD boundaries are correlated with DNA sequences that bind CTCF, which therefore is inferred to block loop extrusion. However, although TADs feature prominently in their Hi-C maps, non-vertebrate eukaryotes either do not express CTCF or show few TAD boundaries that correlate with CTCF sites. In all of these organisms, the counterparts of CTCF remain unknown, frustrating comparisons between Hi-C data and simulations. Results To extend the LEF model across the tree of life, here, we propose the conserved-current loop extrusion (CCLE) model that interprets loop-extruding cohesin as a nearly conserved probability current. From cohesin ChIP-seq data alone, we derive a position-dependent loop extrusion rate, allowing for a modified paradigm for loop extrusion, that goes beyond solely localized barriers to also include loop extrusion rates that vary continuously. We show that CCLE accurately predicts the TAD-scale Hi-C maps of interphase Schizosaccharomyces pombe, as well as those of meiotic and mitotic Saccharomyces cerevisiae, demonstrating its utility in organisms lacking CTCF. Conclusions The success of CCLE in yeasts suggests that loop extrusion by cohesin is indeed the primary mechanism underlying TADs in these systems. CCLE allows us to obtain loop extrusion parameters such as the LEF density and processivity, which compare well to independent estimates.

Abstract Background The safety of CRISPR-based gene editing methods is of the utmost priority in clinical applications. Previous studies have reported that Cas9 cleavage induced frequent aneuploidy in primary human T cells, but whether cleavage-mediated editing of base editors would generate off-target structure variations remains unknown. Here, we investigate the potential off-target structural variations associated with CRISPR/Cas9, ABE, and CBE editing in mouse embryos and primary human T cells by whole-genome sequencing and single-cell RNA-seq analyses. Results The results show that both Cas9 and ABE generate off-target structural variations (SVs) in mouse embryos, while CBE induces rare SVs. In addition, off-target large deletions are detected in 32.74% of primary human T cells transfected with Cas9 and 9.17% of cells transfected with ABE. Moreover, Cas9-induced aneuploid cells activate the P53 and apoptosis pathways, whereas ABE-associated aneuploid cells significantly upregulate cell cycle-related genes and are arrested in the G0 phase. A percentage of 16.59% and 4.29% aneuploid cells are still observable at 3 weeks post transfection of Cas9 or ABE. These off-target phenomena in ABE are universal as observed in other cell types such as B cells and Huh7. Furthermore, the off-target SVs are significantly reduced in cells treated with high-fidelity ABE (ABE-V106W). Conclusions This study shows both CRISPR/Cas9 and ABE induce off-target SVs in mouse embryos and primary human T cells, raising an urgent need for the development of high-fidelity gene editing tools.

Abstract Spatial transcriptomics is a cutting-edge technique that combines gene expression with spatial information, allowing researchers to study molecular patterns within tissue architecture. Here, we present IAMSAM, a user-friendly web-based tool for analyzing spatial transcriptomics data focusing on morphological features. IAMSAM accurately segments tissue images using the Segment Anything Model, allowing for the semi-automatic selection of regions of interest based on morphological signatures. Furthermore, IAMSAM provides downstream analysis, such as identifying differentially expressed genes, enrichment analysis, and cell type prediction within the selected regions. With its simple interface, IAMSAM empowers researchers to explore and interpret heterogeneous tissues in a streamlined manner.

Abstract Unveiling functional relationships between various molecular cell phenotypes from data using machine learning models is a key promise of multiomics. Existing methods either use flexible but hard-to-interpret models or simpler, misspecified models. VI-VS (Variational Inference for Variable Selection) balances flexibility and interpretability to identify relevant feature relationships in multiomic data. It uses deep generative models to identify conditionally dependent features, with false discovery rate control. VI-VS is available as an open-source Python package, providing a robust solution to identify features more likely representing genuine causal relationships.

Abstract Spatial transcriptomics is revolutionizing the exploration of intratissue heterogeneity in cancer, yet capturing cellular niches and their spatial relationships remains challenging. We introduce SpottedPy, a Python package designed to identify tumor hotspots and map spatial interactions within the cancer ecosystem. Using SpottedPy, we examine epithelial-mesenchymal plasticity in breast cancer and highlight stable niches associated with angiogenic and hypoxic regions, shielded by CAFs and macrophages. Hybrid and mesenchymal hotspot distribution follows transformation gradients reflecting progressive immunosuppression. Our method offers flexibility to explore spatial relationships at different scales, from immediate neighbors to broader tissue modules, providing new insights into tumor microenvironment dynamics.

Abstract The tree of blobs of a species network shows only the tree-like aspects of relationships of taxa on a network, omitting information on network substructures where hybridization or other types of lateral transfer of genetic information occur. By isolating such regions of a network, inference of the tree of blobs can serve as a starting point for a more detailed investigation, or indicate the limit of what may be inferrable without additional assumptions. Building on our theoretical work on the identifiability of the tree of blobs from gene quartet distributions under the Network Multispecies Coalescent model, we develop an algorithm, TINNiK, for statistically consistent tree of blobs inference. We provide examples of its application to both simulated and empirical datasets, utilizing an implementation in the MSCquartets 2.0 R package.

Abstract The rapid advancement of spatial transcriptomics technologies has revolutionized our understanding of cell heterogeneity and intricate spatial structures within tissues and organs. However, the high dimensionality and noise in spatial transcriptomic data present significant challenges for downstream data analyses. Here, we develop GraphPCA, an interpretable and quasi-linear dimension reduction algorithm that leverages the strengths of graphical regularization and principal component analysis. Comprehensive evaluations on simulated and multi-resolution spatial transcriptomic datasets generated from various platforms demonstrate the capacity of GraphPCA to enhance downstream analysis tasks including spatial domain detection, denoising, and trajectory inference compared to other state-of-the-art methods.

Abstract Tumours exhibit high genotypic and transcriptional heterogeneity. Both affect cancer progression and treatment, but have been predominantly studied separately in follicular lymphoma. To comprehensively investigate the evolution and genotype-to-phenotype maps in follicular lymphoma, we introduce CaClust, a probabilistic graphical model integrating deep whole exome, single-cell RNA and B-cell receptor sequencing data to infer clone genotypes, cell-to-clone mapping, and single-cell genotyping. CaClust outperforms a state-of-the-art model on simulated and patient data. In-depth analyses of single cells from four samples showcase effects of driver mutations, follicular lymphoma evolution, possible therapeutic targets, and single-cell genotyping that agrees with an independent targeted resequencing experiment.

Abstract Oxford Nanopore Technologies (ONT) offers ultrahigh-throughput multi-sample sequencing but only provides barcode kits that enable up to 96-sample multiplexing. We present TDFPS-Designer, a new toolkit for nanopore sequencing barcode design, which creates significantly more barcodes: 137 with a length of 20 base pairs, 410 at 24 bp, and 1779 at 30 bp, far surpassing ONT’s offerings. It includes GPU-based acceleration for ultra-fast demultiplexing and designs robust barcodes suitable for high-error ONT data. TDFPS-Designer outperforms current methods, improving the demultiplexing recall rate by 20% relative to Guppy, without a reduction in precision.

Abstract The low resolution of spatial transcriptomics data necessitates additional information for optimal use. We developed scDOT, which combines spatial transcriptomics and single cell RNA sequencing to improve the ability to reconstruct single cell resolved spatial maps and identify senescent cells. scDOT integrates optimal transport and expression deconvolution to learn non-linear couplings between cells and spots and to infer cell placements. Application of scDOT to lung spatial transcriptomics data improves on prior methods and allows the identification of the spatial organization of senescent cells, their neighboring cells and novel genes involved in cell-cell interactions that may be driving senescence.

Abstract A recent study reported exaggerated false positives by popular differential expression methods when analyzing large population samples. We reproduce the differential expression analysis simulation results and identify a caveat in the data generation process. Data not truly generated under the null hypothesis led to incorrect comparisons of benchmark methods. We provide corrected simulation results that demonstrate the good performance of dearseq and argue against the superiority of the Wilcoxon rank-sum test as suggested in the previous study.

Abstract A recent study found severely inflated type I error rates for DESeq2 and edgeR, two dominant tools used for differential expression analysis of RNA-seq data. Here, we show that by properly addressing the outliers in the RNA-Seq data using winsorization, the type I error rate of DESeq2 and edgeR can be substantially reduced, and the power is comparable to Wilcoxon rank-sum test for large datasets. Therefore, as an alternative to Wilcoxon rank-sum test, they may still be applied for differential expression analysis of large RNA-Seq datasets.

Abstract Background Transcription factors (TFs) bind to DNA in a highly sequence-specific manner. This specificity manifests itself in vivo as differences in TF occupancy between the two alleles at heterozygous loci. Genome-scale assays such as ChIP-seq currently are limited in their power to detect allele-specific binding (ASB) both in terms of read coverage and representation of individual variants in the cell lines used. This makes prediction of allelic differences in TF binding from sequence alone desirable, provided that the reliability of such predictions can be quantitatively assessed. Results We here propose methods for benchmarking sequence-to-affinity models for TF binding in terms of their ability to predict allelic imbalances in ChIP-seq counts. We use a likelihood function based on an over-dispersed binomial distribution to aggregate evidence for allelic preference across the genome without requiring statistical significance for individual variants. This allows us to systematically compare predictive performance when multiple binding models for the same TF are available. To facilitate the de novo inference of high-quality models from paired-end in vivo binding data such as ChIP-seq, ChIP-exo, and CUT&Tag without read mapping or peak calling, we introduce an extensible reimplementation of our biophysically interpretable machine learning framework named PyProBound. Explicitly accounting for assay-specific bias in DNA fragmentation rate when training on ChIP-seq yields improved TF binding models. Moreover, we show how PyProBound can leverage our threshold-free ASB likelihood function to perform de novo motif discovery using allele-specific ChIP-seq counts. Conclusion Our work provides new strategies for predicting the functional impact of non-coding variants.

Abstract Two correspondences raised concerns or comments about our analyses regarding exaggerated false positives found by differential expression (DE) methods. Here, we discuss the points they raise and explain why we agree or disagree with these points. We add new analysis to confirm that the Wilcoxon rank-sum test remains the most robust method compared to the other five DE methods (DESeq2, edgeR, limma-voom, dearseq, and NOISeq) in two-condition DE analyses after considering normalization and winsorization, the data preprocessing steps discussed in the two correspondences.

Abstract Background MicroRNAs (miRNAs) are small non-coding RNAs that post-transcriptionally regulate gene expression. Perturbations in plasma miRNA levels are known to impact disease risk and have potential as disease biomarkers. Exploring the genetic regulation of miRNAs may yield new insights into their important role in governing gene expression and disease mechanisms. Results We present genome-wide association studies of 2083 plasma circulating miRNAs in 2178 participants of the Rotterdam Study to identify miRNA-expression quantitative trait loci (miR-eQTLs). We identify 3292 associations between 1289 SNPs and 63 miRNAs, of which 65% are replicated in two independent cohorts. We demonstrate that plasma miR-eQTLs co-localise with gene expression, protein, and metabolite-QTLs, which help in identifying miRNA-regulated pathways. We investigate consequences of alteration in circulating miRNA levels on a wide range of clinical conditions in phenome-wide association studies and Mendelian randomisation using the UK Biobank data (N = 423,419), revealing the pleiotropic and causal effects of several miRNAs on various clinical conditions. In the Mendelian randomisation analysis, we find a protective causal effect of miR-1908-5p on the risk of benign colon neoplasm and show that this effect is independent of its host gene (FADS1). Conclusions This study enriches our understanding of the genetic architecture of plasma miRNAs and explores the signatures of miRNAs across a wide range of clinical conditions. The integration of population-based genomics, other omics layers, and clinical data presents opportunities to unravel potential clinical significance of miRNAs and provides tools for novel miRNA-based therapeutic target discovery.

Abstract Spatial transcriptomics technologies have been widely applied to decode cellular distribution by resolving gene expression profiles in tissue. However, sequencing techniques still limit the ability to create a fine-resolved spatial cell-type map. To this end, we develop a novel deep-learning-based approach, STASCAN, to predict the spatial cellular distribution of captured or uncharted areas where only histology images are available by cell feature learning integrating gene expression profiles and histology images. STASCAN is successfully applied across diverse datasets from different spatial transcriptomics technologies and displays significant advantages in deciphering higher-resolution cellular distribution and resolving enhanced organizational structures.

Abstract Numerous plant taxa are threatened by habitat destruction or overexploitation. To overcome these threats, new methods are urgently needed for rescuing threatened and endangered plant species. Here, we review the genetic consequences of threats to species populations. We highlight potential advantages of genome editing for mitigating negative effects caused by new pathogens and pests or climate change where other approaches have failed. We propose solutions to protect threatened plants using genome editing technology unless absolutely necessary. We further discuss the challenges associated with genome editing in plant conservation to mitigate the decline of plant diversity.

Abstract The accurate reconstruction of genome-scale metabolic models (GEMs) for unculturable species poses challenges due to the incomplete and fragmented genetic information typical of metagenome-assembled genomes (MAGs). While existing tools leverage sequence homology from single genomes, this study introduces pan-Draft, a pan-reactome-based approach exploiting recurrent genetic evidence to determine the solid core structure of species-level GEMs. By comparing MAGs clustered at the species-level, pan-Draft addresses the issues due to the incompleteness and contamination of individual genomes, providing high-quality draft models and an accessory reactions catalog supporting the gapfilling step. This approach will improve our comprehension of metabolic functions of uncultured species.

Abstract Simultaneous profiling of single-cell gene expression and lineage history holds enormous potential for studying cellular decision-making. Recent computational approaches combine both modalities into cellular trajectories; however, they cannot make use of all available lineage information in destructive time-series experiments. Here, we present moslin, a Gromov-Wasserstein-based model to couple cellular profiles across time points based on lineage and gene expression information. We validate our approach in simulations and demonstrate on Caenorhabditis elegans embryonic development how moslin predicts fate probabilities and putative decision driver genes. Finally, we use moslin to delineate lineage relationships among transiently activated fibroblast states during zebrafish heart regeneration.

Abstract Background Pervasive translation is a widespread phenomenon that plays a critical role in the emergence of novel microproteins, but the diversity of translation patterns contributing to their generation remains unclear. Based on 54 ribosome profiling (Ribo-Seq) datasets, we investigated the yeast Ribo-Seq landscape using a representation framework that allows the comprehensive inventory and classification of the entire diversity of Ribo-Seq signals, including non-canonical ones. Results We show that if coding regions occupy specific areas of the Ribo-Seq landscape, noncoding regions encompass a wide diversity of Ribo-Seq signals and, conversely, populate the entire landscape. Our results show that pervasive translation can, nevertheless, be associated with high specificity, with 1055 noncoding ORFs exhibiting canonical Ribo-Seq signals. Using mass spectrometry under standard conditions or proteasome inhibition with an in-house analysis protocol, we report 239 microproteins originating from noncoding ORFs that display canonical but also non-canonical Ribo-Seq signals. Each condition yields dozens of additional microprotein candidates with comparable translation properties, suggesting a larger population of volatile microproteins that are challenging to detect. Our findings suggest that non-canonical translation signals may harbor valuable information and underscore the significance of considering them in proteogenomic studies. Finally, we show that the translation outcome of a noncoding ORF is primarily determined by the initiating codon and the codon distribution in its two alternative frames, rather than features indicative of functionality. Conclusion Our results enable us to propose a topology of a species’ Ribo-Seq landscape, opening the way to comparative analyses of this translation landscape under different conditions.

Abstract Single-cell RNA-sequencing (scRNA-seq) provides gene expression profiles of individual cells from complex samples, facilitating the detection of cell type-specific marker genes. In scRNA-seq experiments with multiple donors, the population level variation brings an extra layer of complexity in cell type-specific gene detection, for example, they may not appear in all donors. Motivated by this observation, we develop a statistical model named scCTS to identify cell type-specific genes from population-level scRNA-seq data. Extensive data analyses demonstrate that the proposed method identifies more biologically meaningful cell type-specific genes compared to traditional methods.

Abstract The extremely high levels of genetic polymorphism within the human major histocompatibility complex (MHC) limit the usefulness of reference-based alignment methods for sequence assembly. We incorporate a short-read, de novo assembly algorithm into a workflow for novel application to the MHC. MHConstructor is a containerized pipeline designed for high-throughput, haplotype-informed, reproducible assembly of both whole genome sequencing and target capture short-read data in large, population cohorts. To-date, no other self-contained tool exists for the generation of de novo MHC assemblies from short-read data. MHConstructor facilitates wide-spread access to high-quality, alignment-free MHC sequence analysis.

Abstract The identification of gene regulatory networks (GRNs) is crucial for understanding cellular differentiation. Single-cell RNA sequencing data encode gene-level covariations at high resolution, yet data sparsity and high dimensionality hamper accurate and scalable GRN reconstruction. To overcome these challenges, we introduce NetID leveraging homogenous metacells while avoiding spurious gene–gene correlations. Benchmarking demonstrates superior performance of NetID compared to imputation-based methods. By incorporating cell fate probability information, NetID facilitates the prediction of lineage-specific GRNs and recovers known network motifs governing bone marrow hematopoiesis, making it a powerful toolkit for deciphering gene regulatory control of cellular differentiation from large-scale single-cell transcriptome data.

Abstract Background Lactylation, a post-translational modification, is increasingly recognized for its role in cancer progression. This study investigates its prevalence and impact in oral squamous cell carcinoma (OSCC). Results Immunohistochemical staining of 81 OSCC cases shows lactylation levels correlate with malignancy grading. Proteomic analyses of six OSCC tissue pairs reveal 2765 lactylation sites on 1033 proteins, highlighting its extensive presence. These modifications influence metabolic processes, molecular synthesis, and transport. CAL27 cells are subjected to cleavage under targets and tagmentation assay for accessible-chromatin with high-throughput sequencing, and transcriptomic sequencing pre- and post-lactate treatment, with 217 genes upregulated due to lactylation. Chromatin immunoprecipitation-quantitative PCR and real-time fluorescence quantitative PCR confirm the regulatory role of lactylation at the K146 site of dexh-box helicase 9 (DHX9), a key factor in OSCC progression. CCK8, colony formation, scratch healing, and Transwell assays demonstrate that lactylation mitigates the inhibitory effect of DHX9 on OSCC, thereby promoting its occurrence and development. Conclusions Lactylation actively modulates gene expression in OSCC, with significant effects on chromatin structure and cellular processes. This study provides a foundation for developing targeted therapies against OSCC, leveraging the role of lactylation in disease pathogenesis.

Abstract Spatial barcoding-based transcriptomic (ST) data require deconvolution for cellular-level downstream analysis. Here we present SDePER, a hybrid machine learning and regression method to deconvolve ST data using reference single-cell RNA sequencing (scRNA-seq) data. SDePER tackles platform effects between ST and scRNA-seq data, ensuring a linear relationship between them while addressing sparsity and spatial correlations in cell types across capture spots. SDePER estimates cell-type proportions, enabling enhanced resolution tissue mapping by imputing cell-type compositions and gene expressions at unmeasured locations. Applications to simulated data and four real datasets showed SDePER’s superior accuracy and robustness over existing methods.

Abstract Isobaric labeling-based mass spectrometry (ILMS) has been widely used to quantify, on a proteome-wide scale, the relative protein abundance in different biological conditions. However, large-scale ILMS data sets typically involve multiple runs of mass spectrometry, bringing great computational difficulty to the integration of ILMS samples. We present zMAP, a toolset that makes ILMS intensities comparable across mass spectrometry runs by modeling the associated mean-variance dependence and accordingly applying a variance stabilizing z-transformation. The practical utility of zMAP is demonstrated in several case studies involving the dynamics of cell differentiation and the heterogeneity across cancer patients.

Abstract The exponential increase in sequencing data calls for conceptual and computational advances to extract useful biological insights. One such advance, minimizers, allows for reducing the quantity of data handled while maintaining some of its key properties. We provide a basic introduction to minimizers, cover recent methodological developments, and review the diverse applications of minimizers to analyze genomic data, including de novo genome assembly, metagenomics, read alignment, read correction, and pangenomes. We also touch on alternative data sketching techniques including universal hitting sets, syncmers, or strobemers. Minimizers and their alternatives have rapidly become indispensable tools for handling vast amounts of data.

Abstract Methylation quantitative trait loci (meQTLs) quantify the effects of genetic variants on DNA methylation levels. However, most published studies utilize bulk methylation datasets composed of different cell types and limit our understanding of cell-type-specific methylation regulation. We propose a hierarchical Bayesian interaction (HBI) model to infer cell-type-specific meQTLs, which integrates a large-scale bulk methylation data and a small-scale cell-type-specific methylation data. Through simulations, we show that HBI enhances the estimation of cell-type-specific meQTLs. In real data analyses, we demonstrate that HBI can further improve the functional annotation of genetic variants and identify biologically relevant cell types for complex traits.

Abstract Increasingly, scRNA-Seq studies explore cell populations across different samples and the effect of sample heterogeneity on organism’s phenotype. However, relatively few bioinformatic methods have been developed which adequately address the variation between samples for such population-level analyses. We propose a framework for representing the entire single-cell profile of a sample, which we call a GloScope representation. We implement GloScope on scRNA-Seq datasets from study designs ranging from 12 to over 300 samples and demonstrate how GloScope allows researchers to perform essential bioinformatic tasks at the sample-level, in particular visualization and quality control assessment.

Abstract Background In response to drought stress (DS), plants undergo complex processes that entail significant transcriptome reprogramming. However, the intricate relationship between the dynamic alterations in the three-dimensional (3D) genome and the modulation of gene co-expression in drought responses remains a relatively unexplored area. Results In this study, we reconstruct high-resolution 3D genome maps based on genomic regions marked by H3K9ac, an active histone modification that dynamically responds to soil water variations in rice. We discover a genome-wide disconnection of 3D genome contact upon DS with over 10,000 chromatin loops lost, which are partially recovered in the subsequent re-watering. Loops integrating promoter–promoter interactions (PPI) contribute to gene expression in addition to basal H3K9ac modifications. Moreover, H3K9ac-marked promoter regions with high affinities in mediating PPIs, termed as super-promoter regions (SPRs), integrate spatially clustered PPIs in a super-enhancer-like manner. Interestingly, the knockout mutation of OsbZIP23, a well-defined DS-responsive transcription factor, leads to the disassociation of over 80% DS-specific PPIs and decreased expression of the corresponding genes under DS. As a case study, we show how OsbZIP23 integrates the PPI cluster formation and the co-expression of four dehydrin genes, RAB16A–D, through targeting the RAB16C SPR in a stress signaling-dependent manner. Conclusions Our high-resolution 3D genome maps unveil the principles and details of dynamic genome folding in response to water supply variations and illustrate OsbZIP23 as an indispensable integrator of the yet unique 3D genome organization that is essential for gene co-expression under DS in rice.

Abstract Haplotype information is crucial for biomedical and population genetics research. However, current strategies to produce de novo haplotype-resolved assemblies often require either difficult-to-acquire parental data or an intermediate haplotype-collapsed assembly. Here, we present Graphasing, a workflow which synthesizes the global phase signal of Strand-seq with assembly graph topology to produce chromosome-scale de novo haplotypes for diploid genomes. Graphasing readily integrates with any assembly workflow that both outputs an assembly graph and has a haplotype assembly mode. Graphasing performs comparably to trio phasing in contiguity, phasing accuracy, and assembly quality, outperforms Hi-C in phasing accuracy, and generates human assemblies with over 18 chromosome-spanning haplotypes.

Abstract Background Respiratory diseases impose an immense health burden worldwide. Epidemiological studies have revealed extensive disparities in the incidence and severity of respiratory tract infections between men and women. It has been hypothesized that there might also be a nasal microbiome axis contributing to the observed sex disparities. Results Here, we study the nasal microbiome of healthy young adults in the largest cohort to date with 1593 individuals, using shotgun metagenomic sequencing. We compile the most comprehensive reference catalog for the nasal bacterial community containing 4197 metagenome-assembled genomes and integrate the mycobiome, to provide a valuable resource and a more holistic perspective for the understudied human nasal microbiome. We systematically evaluate sex differences and reveal extensive sex-specific features in both taxonomic and functional levels in the nasal microbiome. Through network analyses, we capture markedly higher ecological stability and antagonistic potentials in the female nasal microbiome compared to the male’s. The analysis of the keystone bacteria reveals that the sex-dependent evolutionary characteristics might have contributed to these differences. Conclusions In summary, we construct the most comprehensive catalog of metagenome-assembled-genomes for the nasal bacterial community to provide a valuable resource for the understudied human nasal microbiome. On top of that, comparative analysis in relative abundance and microbial co-occurrence networks identify extensive sex differences in the respiratory tract community, which may help to further our understanding of the observed sex disparities in the respiratory diseases.

Abstract Indexing techniques relying on k-mers have proven effective in searching for RNA sequences across thousands of RNA-seq libraries, but without enabling direct RNA quantification. We show here that arbitrary RNA sequences can be quantified in seconds through their decomposition into k-mers, with a precision akin to that of conventional RNA quantification methods. Using an index of the Cancer Cell Line Encyclopedia (CCLE) collection consisting of 1019 RNA-seq samples, we show that k-mer indexing offers a powerful means to reveal non-reference sequences, and variant RNAs induced by specific gene alterations, for instance in splicing factors.

Abstract Background Diffuse invasion of glioblastoma cells through normal brain tissue is a key contributor to tumor aggressiveness, resistance to conventional therapies, and dismal prognosis in patients. A deeper understanding of how components of the tumor microenvironment (TME) contribute to overall tumor organization and to programs of invasion may reveal opportunities for improved therapeutic strategies. Results Towards this goal, we apply a novel computational workflow to a spatiotemporally profiled GBM xenograft cohort, leveraging the ability to distinguish human tumor from mouse TME to overcome previous limitations in the analysis of diffuse invasion. Our analytic approach, based on unsupervised deconvolution, performs reference-free discovery of cell types and cell activities within the complete GBM ecosystem. We present a comprehensive catalogue of 15 tumor cell programs set within the spatiotemporal context of 90 mouse brain and TME cell types, cell activities, and anatomic structures. Distinct tumor programs related to invasion align with routes of perivascular, white matter, and parenchymal invasion. Furthermore, sub-modules of genes serving as program network hubs are highly prognostic in GBM patients. Conclusion The compendium of programs presented here provides a basis for rational targeting of tumor and/or TME components. We anticipate that our approach will facilitate an ecosystem-level understanding of the immediate and long-term consequences of such perturbations, including the identification of compensatory programs that will inform improved combinatorial therapies.

Abstract Precise calling of promiscuous adenosine-to-inosine RNA editing sites from transcriptomic datasets is hindered by DNA mutations and sequencing/mapping errors. Here, we present a stepwise computational framework, called DEMINING, to distinguish RNA editing and DNA mutations directly from RNA sequencing datasets, with an embedded deep learning model named DeepDDR. After transfer learning, DEMINING can also classify RNA editing sites and DNA mutations from non-primate sequencing samples. When applied in samples from acute myeloid leukemia patients, DEMINING uncovers previously underappreciated DNA mutation and RNA editing sites; some associated with the upregulated expression of host genes or the production of neoantigens.

Abstract Reconstructing premortem DNA methylation levels in ancient DNA has led to breakthrough studies such as the prediction of anatomical features of the Denisovan. These studies rely on computationally inferring methylation levels from damage signals in naturally deaminated cytosines, which requires expensive high-coverage genomes. Here, we test two methods for direct methylation measurement developed for modern DNA based on either bisulfite or enzymatic methylation treatments. Bisulfite treatment shows the least reduction in DNA yields as well as the least biases during methylation conversion, demonstrating that this method can be successfully applied to ancient DNA.

Abstract Background Polygenic risk score (PRS) is a major research topic in human genetics. However, a significant gap exists between PRS methodology and applications in practice due to often unavailable individual-level data for various PRS tasks including model fine-tuning, benchmarking, and ensemble learning. Results We introduce an innovative statistical framework to optimize and benchmark PRS models using summary statistics of genome-wide association studies. This framework builds upon our previous work and can fine-tune virtually all existing PRS models while accounting for linkage disequilibrium. In addition, we provide an ensemble learning strategy named PUMAS-ensemble to combine multiple PRS models into an ensemble score without requiring external data for model fitting. Through extensive simulations and analysis of many complex traits in the UK Biobank, we demonstrate that this approach closely approximates gold-standard analytical strategies based on external validation, and substantially outperforms state-of-the-art PRS methods. Conclusions Our method is a powerful and general modeling technique that can continue to combine the best-performing PRS methods out there through ensemble learning and could become an integral component for all future PRS applications.

Abstract Background Genetic perturbation screens with single-cell readouts have enabled rich phenotyping of gene function and regulatory networks. These approaches have been challenging in vivo, especially in adult disease models such as cancer, which include mixtures of malignant and microenvironment cells. Glioblastoma (GBM) is a fatal cancer, and methods of systematically interrogating gene function and therapeutic targets in vivo, especially in combination with standard of care treatment such as radiotherapy, are lacking. Results Here, we iteratively develop a multiplex in vivo perturb-seq CRISPRi platform for single-cell genetic screens in cancer and tumor microenvironment cells that leverages intracranial convection enhanced delivery of sgRNA libraries into mouse models of GBM. Our platform enables potent silencing of drivers of in vivo growth and tumor maintenance as well as genes that sensitize GBM to radiotherapy. We find radiotherapy rewires transcriptional responses to genetic perturbations in an in vivo-dependent manner, revealing heterogenous patterns of treatment sensitization or resistance in GBM. Furthermore, we demonstrate targeting of genes that function in the tumor microenvironment, enabling alterations of ligand-receptor interactions between immune and stromal cells following in vivo CRISPRi perturbations that can affect tumor cell phagocytosis. Conclusion In sum, we demonstrate the utility of multiplexed perturb-seq for in vivo single-cell dissection of adult cancer and normal tissue biology across multiple cell types in the context of therapeutic intervention, a platform with potential for broad application.

Abstract Background Methylation of histone 3 lysine 36 (H3K36me) has emerged as an essential epigenetic component for the faithful regulation of gene expression. Despite its importance in development and disease, how the molecular agents collectively shape the H3K36me landscape is unclear. Results We use mouse mesenchymal stem cells to perturb the H3K36me methyltransferases (K36MTs) and infer the activities of the five most prominent enzymes: SETD2, NSD1, NSD2, NSD3, and ASH1L. We find that H3K36me2 is the most abundant of the three methylation states and is predominantly deposited at intergenic regions by NSD1, and partly by NSD2. In contrast, H3K36me1/3 are most abundant within exons and are positively correlated with gene expression. We demonstrate that while SETD2 deposits most H3K36me3, it may also deposit H3K36me2 within transcribed genes. Additionally, loss of SETD2 results in an increase of exonic H3K36me1, suggesting other (K36MTs) prime gene bodies with lower methylation states ahead of transcription. While NSD1/2 establish broad intergenic H3K36me2 domains, NSD3 deposits H3K36me2 peaks on active promoters and enhancers. Meanwhile, the activity of ASH1L is restricted to the regulatory elements of developmentally relevant genes, and our analyses implicate PBX2 as a potential recruitment factor. Conclusions Within genes, SETD2 primarily deposits H3K36me3, while the other K36MTs deposit H3K36me1/2 independently of SETD2 activity. For the deposition of H3K36me1/2, we find a hierarchy of K36MT activities where NSD1 > NSD2 > NSD3 > ASH1L. While NSD1 and NSD2 are responsible for most genome-wide propagation of H3K36me2, the activities of NSD3 and ASH1L are confined to active regulatory elements.

Abstract Background Alternative polyadenylation (APA) affects most human genes and is recurrently dysregulated in all studied cancers. However, the mechanistic origins of this dysregulation are incompletely understood. Results We describe an unbiased analysis of molecular regulators of poly(A) site selection across The Cancer Genome Atlas and identify that colorectal adenocarcinoma is an outlier relative to all other cancer subtypes. This distinction arises from the frequent presence of loss-of-function APC mutations in colorectal adenocarcinoma, which are strongly associated with long 3′ UTR expression relative to tumors lacking APC mutations. APC knockout similarly dysregulates APA in human colon organoids. By mining previously published APC eCLIP data, we show that APC preferentially binds G- and C-rich motifs just upstream of proximal poly(A) sites. Lastly, we find that reduced APC expression is associated with APA dysregulation in tumor types lacking recurrent APC mutations. Conclusions As APC has been previously identified as an RNA-binding protein that preferentially binds 3′ UTRs during mouse neurogenesis, our results suggest that APC promotes proximal poly(A) site use and that APC loss and altered expression contribute to pervasive APA dysregulation in cancers.

Abstract The B cell lineage tree encapsulates the successive phases of B cell differentiation and maturation, transitioning from hematopoietic stem cells to mature, antibody-secreting cells within the immune system. Mathematically, this lineage can be conceptualized as an evolutionary tree, where each node represents a distinct stage in B cell development, and the edges reflect the differentiation pathways. To compare these lineage trees, a rigorous mathematical metric is essential. Analyzing B cell lineage trees mathematically and quantifying changes in lineage attributes over time necessitates a comparison methodology capable of accurately assessing and measuring these changes. Addressing the intricacies of multiple B cell lineage tree comparisons, this study introduces a novel metric that enhances the precision of comparative analysis. This metric is formulated on principles of metric theory and evolutionary biology, quantifying the dissimilarities between lineage trees by measuring branch length distance and weight. By providing a framework for systematically classifying lineage trees, this metric facilitates the development of predictive models that are crucial for the creation of targeted immunotherapy and vaccines. To validate the effectiveness of this new metric, synthetic datasets that mimic the complexity and variability of real B cell lineage structures are employed. We demonstrated the ability of the new metric method to accurately capture the evolutionary nuances of B cell lineages.

Abstract Background Embryos are regeneration and wound healing masters. They rapidly close wounds and scarlessly remodel and regenerate injured tissue. Regeneration has been extensively studied in many animal models using new tools such as single-cell analysis. However, until now, they have been based primarily on experiments assessing from 1 day post injury. Results In this paper, we reveal that critical steps initiating regeneration occur within hours after injury. We discovered the regeneration initiating cells (RICs) using single-cell and spatial transcriptomics of the regenerating Xenopus laevis tail. RICs are formed transiently from the basal epidermal cells, and their expression signature suggests they are important for modifying the surrounding extracellular matrix thus regulating development. The absence or deregulation of RICs leads to excessive extracellular matrix deposition and defective regeneration. Conclusion RICs represent a newly discovered transient cell state involved in the initiation of the regeneration process.

Abstract Background Powdery mildew is a disease with one of the most substantial impacts on cucumber production globally. The most efficient approach for controlling powdery mildew is the development of genetic resistance; however, few genes associated with inherent variations in cucumber powdery mildew resistance have been identified as of yet. Results In this study, we re-sequence 299 cucumber accessions, which are divided into four geographical groups. A genome-wide association study identifies 50 sites significantly associated with natural variations in powdery mildew resistance. Linkage disequilibrium analysis further divides these 50 sites into 32 linkage disequilibrium blocks containing 41 putative genes. Virus-induced gene silencing and gene expression analysis implicate CsGy5G015960, which encodes a phosphate transporter, as the candidate gene regulating powdery mildew resistance. On the basis of the resequencing data, we generate five CsGy5G015960 haplotypes, identifying Hap.1 as the haplotype most likely associated with powdery mildew resistance. In addition, we determine that a 29-bp InDel in the 3′ untranslated region of CsGy5G015960 is responsible for mRNA stability. Overexpression of CsGy5G015960 Hap.1 in the susceptible line enhances powdery mildew resistance and phosphorus accumulation. Further comparative RNA-seq analysis demonstrates that CsGy5G015960 Hap.1 may regulate cucumber powdery mildew resistance by maintaining a higher H2O2 level through the depletion of multiple class III peroxidases. Conclusions Here we identify a candidate powdery mildew-resistant gene in cucumber using GWAS. The identified gene may be a promising target for molecular breeding and genetic engineering in cucumber to enhance powdery mildew resistance.

Abstract Batch effects in omics data are notoriously common technical variations unrelated to study objectives, and may result in misleading outcomes if uncorrected, or hinder biomedical discovery if over-corrected. Assessing and mitigating batch effects is crucial for ensuring the reliability and reproducibility of omics data and minimizing the impact of technical variations on biological interpretation. In this review, we highlight the profound negative impact of batch effects and the urgent need to address this challenging problem in large-scale omics studies. We summarize potential sources of batch effects, current progress in evaluating and correcting them, and consortium efforts aiming to tackle them.

Abstract In this work, we extend vcfdist to be the first variant call benchmarking tool to jointly evaluate phased single-nucleotide polymorphisms (SNPs), small insertions/deletions (INDELs), and structural variants (SVs) for the whole genome. First, we find that a joint evaluation of small and structural variants uniformly reduces measured errors for SNPs (− 28.9%), INDELs (− 19.3%), and SVs (− 52.4%) across three datasets. vcfdist also corrects a common flaw in phasing evaluations, reducing measured flip errors by over 50%. Lastly, we show that vcfdist is more accurate than previously published works and on par with the newest approaches while providing improved result interpretability.

Abstract Background The shoot apical meristem (SAM), from which all above-ground tissues of plants are derived, is critical to plant morphology and development. In maize (Zea mays), loss-of-function mutant studies have identified several SAM-related genes, most encoding homeobox transcription factors (TFs), located upstream of hierarchical networks of hundreds of genes. Results Here, we collect 46 transcriptome and 16 translatome datasets across 62 different tissues or stages from the maize inbred line B73. We construct a dynamic regulome for 27 members of three SAM-related homeobox subfamilies (KNOX, WOX, and ZF-HD) through machine-learning models for the detection of TF targets across different tissues and stages by combining tsCUT&Tag, ATAC-seq, and expression profiling. This dynamic regulome demonstrates the distinct binding specificity and co-factors for these homeobox subfamilies, indicative of functional divergence between and within them. Furthermore, we assemble a SAM dynamic regulome, illustrating potential functional mechanisms associated with plant architecture. Lastly, we generate a wox13a mutant that provides evidence that WOX13A directly regulates Gn1 expression to modulate plant height, validating the regulome of SAM-related homeobox genes. Conclusions The SAM-related homeobox transcription-factor regulome presents an unprecedented opportunity to dissect the molecular mechanisms governing SAM maintenance and development, thereby advancing our understanding of maize growth and shoot architecture.

Abstract Background Telomeric repeat arrays at the ends of chromosomes are highly dynamic in composition, but their repetitive nature and technological limitations have made it difficult to assess their true variation in genome diversity surveys. Results We have comprehensively characterized the sequence variation immediately adjacent to the canonical telomeric repeat arrays at the very ends of chromosomes in 74 genetically diverse Arabidopsis thaliana accessions. We first describe several types of distinct telomeric repeat units and then identify evolutionary processes such as local homogenization and higher-order repeat formation that shape diversity of chromosome ends. By comparing largely isogenic samples, we also determine repeat number variation of the degenerate and variant telomeric repeat array at both the germline and somatic levels. Finally, our analysis of haplotype structure uncovers chromosome end-specific patterns in the distribution of variant telomeric repeats, and their linkage to the more proximal non-coding region. Conclusions Our findings illustrate the spectrum of telomeric repeat variation at multiple levels in A. thaliana—in germline and soma, across all chromosome ends, and across genetic groups—thereby expanding our knowledge of the evolution of chromosome ends.

Abstract Background N6-methyladenosine (m6A), the most abundant internal modification on eukaryotic mRNA, and N6, 2′-O-dimethyladenosine (m6Am), are epitranscriptomic marks that function in multiple aspects of posttranscriptional regulation. Fat mass and obesity-associated protein (FTO) can remove both m6A and m6Am; however, little is known about how FTO achieves its substrate selectivity. Results Here, we demonstrate that ZBTB48, a C2H2-zinc finger protein that functions in telomere maintenance, associates with FTO and binds both mRNA and the telomere-associated regulatory RNA TERRA to regulate the functional interactions of FTO with target transcripts. Specifically, depletion of ZBTB48 affects targeting of FTO to sites of m6A/m6Am modification, changes cellular m6A/m6Am levels and, consequently, alters decay rates of target RNAs. ZBTB48 ablation also accelerates growth of HCT-116 colorectal cancer cells and modulates FTO-dependent regulation of Metastasis-associated protein 1 (MTA1) transcripts by controlling the binding to MTA1 mRNA of the m6A reader IGF2BP2. Conclusions Our findings thus uncover a previously unknown mechanism of posttranscriptional regulation in which ZBTB48 co-ordinates RNA-binding of the m6A/m6Am demethylase FTO to control expression of its target RNAs.

Abstract The process of splicing messenger RNA to remove introns plays a central role in creating genes and gene variants. We describe Splam, a novel method for predicting splice junctions in DNA using deep residual convolutional neural networks. Unlike previous models, Splam looks at a 400-base-pair window flanking each splice site, reflecting the biological splicing process that relies primarily on signals within this window. Splam also trains on donor and acceptor pairs together, mirroring how the splicing machinery recognizes both ends of each intron. Compared to SpliceAI, Splam is consistently more accurate, achieving 96% accuracy in predicting human splice junctions.

Abstract Clustering is widely used for single-cell analysis, but current methods are limited in accuracy, robustness, ease of use, and interpretability. To address these limitations, we developed an ensemble clustering method that outperforms other methods at hard clustering without the need for hyperparameter tuning. It also performs soft clustering to characterize continuum-like regions and quantify clustering uncertainty, demonstrated here by mapping the connectivity and intermediate transitions between MNIST handwritten digits and between hypothalamic tanycyte subpopulations. This hyperparameter-randomized ensemble approach improves the accuracy, robustness, ease of use, and interpretability of single-cell clustering, and may prove useful in other fields as well.