Biology subjects

Lyu, R.

Publications and source records attributed to Lyu, R..

17 recordsLinked to original sources

bioRxiv2026

A One-Step Chemoselective Strategy for Hydroxymethylcytosine Sequencing in DNA and RNA

5-Hydroxymethylcytosine (5hmC) is a relatively stable chemical mark formed by oxidation of 5-methylcytosine in DNA, while its RNA counterparts, 5-hydroxymethylcytidine (hm5C) and 2'-O-methyl-5-hydroxymethylcytidine (hm5Cm), have recently emerged as potential regulators of post-transcriptional processes. Existing 5hmC enrichment methods rely on anti-bodies or enzymatic {beta}-glucosyltransferase labeling, are largely DNA-specific, and are not readily applicable to RNA. Here we report bisulfite-assisted labeling by thiol sequencing (BALT-seq), a one-step chemoselective strategy that installs a biotin-thioether handle directly onto hydroxymethylcytosine in DNA or RNA under mild acidic bisulfite conditions, without using enzymes. Mechanistic studies show that thiol capture of the bisulfite-activated 5hmC intermediate competes with canonical cytosine-5-methylenesulfonate formation, yielding a stable adduct that DNA polymerases bypass efficiently and that resists APOBEC-mediated deamination. In mouse embryonic stem cells, BALT-seq produces genome-wide 5hmC profiles highly concordant with 5hmC-Seal, TAB-seq, and oxBS-seq, recovers expected gene-body and chromatin-state enrichment, and extends to low-input cell-free DNA. On RNA, BALT-seq enriches hm5C/hm5Cm-containing transcripts more effectively than antibody-based hMeRIP and reveals structured modification landscapes across tRNA species and LINE/LTR repeat families. BALT-seq thus establishes a unified, enzyme-free chemical platform for hydroxymethylcytosine profiling across DNA, cell-free DNA, and RNA.

genomics↗

bioRxiv2026

A U1-U3 snRNA-snoRNA interaction couples SF3B1 mutation to chromatin-state rewiring and genome instability

Mutations in spliceosome factors such as SF3B1 are recurrent across human diseases, including myelodysplastic syndromes and leukemia1-4, yet splicing defects alone do not fully explain the widespread chromatin alterations and genome instability in mutant cells5. Here, by comprehensively mapping snRNA-directed RNA-RNA interactions, we identify two previously unrecognized interaction motifs in U1 snRNA beyond canonical 5' splice-site pairing6,7. These motifs enable U1 RNA to i) bind intronic and other chromatin-associated RNA (caRNA) regions outside of splice sites, and ii) base pair specifically with snoRNA. We uncover a U1-U3 snRNA-snoRNA interaction that recruits the H3K36 methyltransferase SETD2 to caRNA, promoting gene-body H3K36me3 and antagonizing H3K27me3 to modulate chromatin accessibility. The snRNA-snoRNA interface is essential for this previously unrecognized layer of chromatin and transcriptional regulation mediated through SETD2. SF3B1 mutation enhances U1-U3 binding and increases the association of the U1-U3 complex with caRNA, driving chromatin-accessibility remolding, R-loop formation, DNA damage, and copy-number abnormalities that promote tumorigenesis. A U1-specific 2'-O-methoxyethyl antisense oligonucleotide that selectively blocks U1-U3 pairing suppresses these genomic abnormalities, reduces leukemic infiltration, and prolongs survival in xenograft and patient-derived models, establishing pathological snRNA-snoRNA rewiring as a critical driver of SF3B1-mutant leukemogenesis.

cell biology↗

bioRxiv2025

Ultra-mild bisulfite outperforms EM-seq for 5-methylcytosine detection with low input DNA

We present Ultra-Mild Bisulfite Sequencing (UMBS-seq), a method for 5-methylcytosine (5mC) detection that minimizes DNA degradation and background noise. UMBS-seq outperforms conventional bisulfite and enzymatic methyl-sequencing (EM-seq) methods in library yield, complexity, and conversion efficiency when applied to low-input DNA samples. In particular, its effectiveness with low-input cell-free DNA (cfDNA) and hybridization-based target capture highlight its potential for clinical applications, including 5mC biomarker detection and early disease diagnosis.

genomics↗

bioRxiv2025

ASAP-ID: Proximity labelling with small tags

Biotinylation-based proximity labelling methods are valuable for discovering protein-protein interactions within cellular systems. However, one limitation of these approaches is that most require fusing the target protein with the enzyme that biotinylates nearby proteins (i.e., TurboID or APEX2), which risks sterically disrupting the proteins native function. Here, we present a method designed to reduce the steric impact of these fusions and offer greater flexibility in labelling modalities. The method, Antibody and Small-tag Assembly on Proteins for Interaction Detection (ASAP-ID), involves a bipartite system. Target proteins are fused to a small peptide antigen that recruits TurboID or APEX2 fused to an antibody directed to the antigen. Using two different antigen/antibody systems (SunTag and MoonTag), we show that ASAP-ID can specifically label human Lamin A in cells. The method works when the target protein and nanobody are co-expressed together in cis (ASAP-IDIC). We also demonstrate that the approach works when the antibody fusion is added in trans to fixed cells post-expression (ASAP-IDIT). ASAP-IDIT identified more than 448 known and previously undescribed interactors of lamin. We further used ASAP-IDIT to study how ALS-mutant profilin 1 affected its interactome. The method identified proteins involved in protein quality control that correlated with aggregation propensity. Moreover, the different mutants showed variation in the cellular location where aggregates formed. ASAP-IDIT revealed preferences for mitochondrial proteins for the two profilin mutants that tend to aggregate in the cytoplasm, C71G and M114T, and nuclear proteins for a mutant more prone to nuclear aggregation. These findings position ASAP-ID as a powerful addition to the proximity labelling toolkit, capable of probing subtle differences in interactomes in a less invasive manner.

biochemistry↗

bioRxiv2025

A spatial transcriptomic atlas of acute neonatal lung injury across development and disease severity

A molecular understanding of lung organogenesis requires delineation of the timing and regulation of the cellular transitions that ultimately form and support a surface capable of gas exchange. While the advent of single-cell transcriptomics has allowed for the discovery and identification of transcriptionally distinct cell populations present during lung development, the spatiotemporal dynamics of these transcriptional shifts remain undefined. With imaging-based spatial transcriptomics, we analyzed the gene expression patterns in 17 human infant lungs at varying stages of development and injury, creating a spatial transcriptomic atlas of [~]1.2 million cells. We applied computational clustering approaches to identify shared molecular patterns among this cohort, informing how tissue architecture and molecular spatial relationships are coordinated during development and disrupted in disease. Recognizing that all preterm birth represents an injury to the developing lung, we created a simplified classification scheme that relies upon the routinely collected objective measures of gestational age and life span. Within this framework, we have identified cell type patterns across gestational age and life span variables that would likely be overlooked when using the conventional "disease vs. control" binary comparison. Together, these data represent an open resource for the lung research community, supporting discovery-based inquiry and identification of targetable molecular mechanisms in both normal and arrested human lung development.

genomics↗

bioRxiv2025

SpatialRNA: a Python package for easy application of Graph Neural Network models on single-molecule spatial transcriptomicsdataset

Image-based spatial transcriptomics deliver gene expression measurements of RNA transcripts in tissue slices with single-molecule resolution and spatial context preserved. Modern Graph Neural Network (GNN) models are promising methods for capturing the complex molecular and cellular phenotypes in tissues at single-transcript and single-cell levels. A key application of GNNs is the detection of spatial domains or niches, that is, groups of molecules and/or cells that collaboratively work together to produce complex phenotypes. Due to the vast number of detected transcripts in image-based spatial transcriptomics, applying GNNs on RNA molecule graphs is not trivial. We present a Python package, SpatialRNA, for easy (sub)graph generation from tissue samples and provide comprehensive tutorials for convenient and efficient application of Graph Neural Network models under the PyG framework. Availability and implementation: The SpatialRNA package is freely accessible from GitLab https://gitlab.svi.edu.au/biocellgen-public/spatialrna and can be installed via pip. Python notebooks and scripts used in case studies are freely accessible at https://gitlab.svi.edu.au/biocellgen-public/case_study_ipf and https://gitlab.svi.edu.au/biocellgen-public/case_study_xenium_5k_panel.

bioinformatics↗

bioRxiv2025

Deep-learning-based interpolation of longitudinal microbiome data powers biologically informative discovery

The human microbiome is a foundational and dynamic foundation for several health-related functions and disease processes. Advances in microbiome sequencing have enabled the characterization of microbial communities in several niches. Longitudinal microbiome studies further strive to discover clinically informative microbial community trajectories. However, these data are fraught with dropout events, high noise, and irregular sampling that limit and prevent the use of many available longitudinal analysis tools. To address these challenges, we introduce Bidirectional GRU-ODE-Bayes (BGOB), a deep learning framework developed for longitudinal microbiome interpolation. BGOB combines bidirectional information flow and ODE-based continuous modeling to jointly interpolate and smoothen trends across individual participants, providing uniform, denoised time intervals across patients. BGOB enables vastly improved performance in differential abundance testing and time-to-event analysis, and makes possible longitudinal analyses requiring uniformity, such as lead-lag detection and temporal clustering. After interpolation, previously low-powered datasets are able to broadly recapitulate known microbiology and elucidate interacting microbial communities. We highlight several associations between microbial taxa and disease, including novel species associated with Early Childhood Caries and disruption of key healthy gut microbiota in Inflammatory Bowel Disease. The BGOB package is publicly available at https://github.com/Rachel-Lyu/BGOB_n_test.

microbiology↗

bioRxiv2024

Scalable identification of lineage-specific gene regulatory networks from metacells with NetID

The identification of gene regulatory networks (GRN) governing distinct cell fates in multilineage cellular differentiation systems is of critical importance for understanding cell fate decision. Single-cell RNA-sequencing (scRNA-seq) provides a powerful tool for the quantification of gene-level co-variation across the cell state manifold. However, accurate GRN reconstruction is hampered by the sparsity of scRNA-seq data introducing substantial technical noise. Moreover, the high dimensionality of typical scRNA-seq datasets limits the scalability of available approaches. To overcome these challenges, and to facilitate the inference of lineage-specific GRNs with directed regulator-target relations, we introduce NetID. This approach optimizes coverage of the cell state manifold by homogenous metacells and avoids spurious gene-gene correlations observed with available imputation methods. Benchmarking demonstrates superior performance of NetID compared to imputation-based GRN inference. By incorporating cell fate probability information, NetID facilitates prediction of lineage-specific GRNs and recovers known network motifs centered around lineage-determining transcription factors governing bone marrow hematopoiesis, making it a powerful toolkit for deciphering the gene regulatory control of cellular differentiation from large-scale single-cell transcriptome data.

bioinformatics↗

bioRxiv2024

Quantitative analysis of cis-regulatory elements in transcription with KAS-ATAC-seq

Cis-regulatory elements (CREs) are pivotal in orchestrating gene expression throughout diverse biological systems. Accurate identification and in-depth characterization of functional CREs are crucial for decoding gene regulation network and dynamics during cellular processes. In this study, we developed Kethoxal-Assisted Single-stranded DNA Assay for Transposase-Accessible Chromatin with Sequencing (KAS-ATAC-seq) to provide quantitative insights into transcriptional activity of CREs. A main advantage of KAS-ATAC-seq lies in its precise measurement of ssDNA levels within both proximal and distal ATAC-seq peaks, enabling the identification of transcriptional regulatory sequences in genomes. This feature is particularly adept at defining Single-Stranded Transcribing Enhancers (SSTEs). SSTEs are highly enriched with nascent RNA transcription and specific transcription factors (TFs) binding sites that determine cellular identity. Moreover, KAS-ATAC-seq provides a detailed characterization and functional implications of various SSTE subtypes; KAS-ATAC-seq signals on SSTEs exhibit more robust correlation with enhancer activities when compared with ATAC-seq data and active histone mark profiles. Our analysis of promoters and SSTEs during mouse neural differentiation demonstrates that KAS-ATAC-seq can effectively identify immediate-early activated CREs in response to retinoic acid (RA) treatment. We further discovered that ETS TFs and YY1 are critical in initiating early neural differentiation from mESCs to NPCs. Our findings indicate that KAS-ATAC-seq provides more precise annotation of functional CREs in transcription. Future applications of KAS-ATAC-seq would help elucidate the intricate dynamics of gene regulation in diverse biological processes and biomedical applications.

genomics↗

bioRxiv2023

Image-based spatial transcriptomics identifies molecular niche dysregulation associated with distal lung remodeling in pulmonary fibrosis

The human lung is structurally complex, with a diversity of specialized epithelial, stromal and immune cells playing specific functional roles in anatomically distinct locations, and large-scale changes in the structure and cellular makeup of this distal lung is a hallmark of pulmonary fibrosis (PF) and other progressive chronic lung diseases. Single-cell transcriptomic studies have revealed numerous disease-emergent/enriched cell types/states in PF lungs, but the spatial contexts wherein these cells contribute to disease pathogenesis has remained uncertain. Using sub-cellular resolution image-based spatial transcriptomics, we analyzed the gene expression of more than 1 million cells from 19 unique lungs. Through complementary cell-based and innovative cell-agnostic analyses, we characterized the localization of PF-emergent cell-types, established the cellular and molecular basis of classical PF histopathologic disease features, and identified a diversity of distinct molecularly-defined spatial niches in control and PF lungs. Using machine-learning and trajectory analysis methods to segment and rank airspaces on a gradient from normal to most severely remodeled, we identified a sequence of compositional and molecular changes that associate with progressive distal lung pathology, beginning with alveolar epithelial dysregulation and culminating with changes in macrophage polarization. Together, these results provide a unique, spatially-resolved characterization of the cellular and molecular programs of PF and control lungs, provide new insights into the heterogeneous pathobiology of PF, and establish analytical approaches which should be broadly applicable to other imaging-based spatial transcriptomic studies.

genomics↗

bioRxiv2023

PTPN2 copper-sensing rapidly relays copper level fluctuations into EGFR/CREB activation and associated CTR1 transcriptional repression

Fluxes in human intra- and extracellular copper levels recently garnered attention for roles in cellular signaling, including affecting levels of the signaling molecule cyclic adenosine monophosphate (cAMP). We herein applied an unbiased temporal evaluation of the whole-genome transcriptional activities modulated by fluctuations in copper levels to identify the copper sensor proteins responsible for driving these activities. We found that fluctuations in physiologically-relevant copper levels rapidly modulate EGFR/MAPK/ERK signal transduction and activation of the transcription factor cAMP response element-binding protein (CREB). Both intracellular and extracellular assays support Cu1+ inhibition of the EGFR-phosphatase PTPN2 (and potentially the homologous PTPN1)-via direct ligation to the PTPN2 active site cysteine side chain-as the underlying mechanism of copper-stimulated EGFR signal transduction activation. Depletion of copper represses this signaling pathway. We additionally show i) copper supplementation drives transcriptional repression of the copper importer CTR1 and ii) CREB activity is inversely correlated with CTR1 expression. In summary, our study reveals PTPN2 as a physiological copper sensor and defines a regulatory mechanism linking feedback control of copper-stimulated MAPK/ERK/CREB-signaling and CTR1 expression, thereby uncovering a previously unrecognized link between copper levels and cellular signal transduction.

biochemistry↗

bioRxiv2023

5-hydroxymethylcytosine profiling of cell-free DNA identifies bivalent genes that are prognostic of survival in high-risk neuroblastoma.

Neuroblastoma is the most common extra-cranial solid tumor in childhood and epigenetic dysregulation is a key driver of this embryonal disease. In cell-free DNA from neuroblastoma patients with high-risk disease, we found increased 5-hydroxymethylcytosine (5-hmC) deposition on Polycomb Repressive Complex 2 (PRC2) target genes, a finding previously described in the context of bivalent genes. As bivalent genes, defined as genes bearing both activating (H3K4me3) and repressive (H3K27me3) chromatin modifications, have been shown to play an important role in development and cancer, we investigated the potential role of bivalent genes in maintaining a de-differentiated state in neuroblastoma and their potential use as a biomarker. We identified 313 genes that bore bivalent chromatin marks, were enriched for mediators of neuronal differentiation, and were transcriptionally repressed across a panel of heterogenous neuroblastoma cell lines. Through gene set variance analysis, we developed a clinically implementable bivalent signature. In three distinct clinical cohorts, low bivalent signature was significantly and independently associated with worse clinical outcome in high-risk neuroblastoma patients. Thus, low expression of bivalent genes is a biomarker of ultra-high-risk disease and may represent a therapeutic opportunity in neuroblastoma.

genomics↗

bioRxiv2022

KAS-pipe2: a flexible toolkit for exploring KAS-seq and spKAS-seq data

Kethoxal-assisted ssDNA sequencing (KAS-seq) is gaining popularity as a robust and effective approach to study the dynamics of transcriptionally engaged RNA polymerases through profiling of genome-wide single-stranded DNA (ssDNA). Its latest variant, spKAS-seq, a strand-specific version of KAS-seq, has been developed to map genome-wide R-loop structures by detecting imbalances of ssDNA on two strands. However, user-friendly, open-source analysis pipelines for KAS-seq data are still lacking. Here we present KAS-pipe2 as a flexible and integrated toolkit to facilitate the analysis and interpretation of KAS-seq data. KAS-pipe2 can perform standard analyses such as quality control, read alignment, and differential RNA polymerase activity analysis. In addition, KAS-pipe2 introduces many novel features, including, but not limited to: calculation of transcriptional indexes, identification of single-stranded transcribing enhancers, and high-resolution mapping of R-loops. We use benchmark datasets to demonstrate that KAS-pipe2 provides a powerful framework to study transient transcriptional regulatory programs. KAS-pipe2 is available at https://github.com/Ruitulyu/KAS-pipe2.

bioinformatics↗

bioRxiv2022

Fancm regulates meiotic double-strand break repair pathway choice in mammals

Meiotic crossovers are required for accurate chromosome segregation and to produce new allelic combinations. Meiotic crossover numbers are tightly regulated within a narrow range, despite an excess of initiating DNA double-strand breaks. Here, we describe the tumour suppressor FANCM as a meiotic anti-crossover factor in mammals. Crossover analyses with single-gamete and pedigree datasets both reveal a genome-wide increase in crossover frequencies in Fancm-deficient mice. Gametogenesis is heavily perturbed in Fancm loss of function mice, which is consistent with the reproductive defects reported in humans with biallelic FANCM mutations. A portion of the gametogenesis defects can be attributed to the cGAS-STING pathway. Despite the gametogenesis phenotypes in Fancm mutants both sexes were capable of producing offspring. We propose that the anti-crossover function and role in gametogenesis of Fancm are separable and will inform diagnostic pathways for human genomic instability disorders.

genetics↗

bioRxiv2022

sgcocaller and comapr: personalised haplotype assembly and comparative crossover map analysis using single gametes

Profiling gametes of an individual enables the construction of personalised haplotypes and meiotic crossover landscapes, now achievable at larger scale than ever through the availability of high-throughput single-cell sequencing technologies. However, high-throughput single-gamete data commonly have low depth of coverage per gamete, which challenges existing gametebased haplotype phasing methods. In addition, haplotyping a large number of single gametes from high-throughput singlecell DNA sequencing data and constructing meiotic crossover profiles using existing methods requires intensive processing. Here, we introduce efficient software tools for the essential tasks of generating personalised haplotypes and calling crossovers in gametes from single-gamete DNA sequencing data (sgcocaller), and constructing, visualising, and comparing individualised crossover landscapes from single gametes (comapr). With additional data pre-possessing, the tools can also be applied to bulk-sequenced samples. We demonstrate that sgcocaller is able to generate impeccable phasing results for high-coverage datasets, on which it is more accurate and stable than existing methods, and also performs well on low-coverage single-gamete sequencing datasets for which current methods fail. Our tools achieve highly accurate results with user-friendly installation, comprehensive documentation, efficient computation times and minimal memory usage.

bioinformatics↗

bioRxiv2021

A phylotranscriptome study using silica gel-dried leaf tissues produces an updated robust phylogeny of Ranunculaceae

The utility of transcriptome data in plant phylogenetics has gained popularity in recent years. However, because RNA degrades much more easily than DNA, the logistics of obtaining fresh tissues has become a major limiting factor for widely applying this method. Here, we used Ranunculaceae to test whether silica-dried plant tissues could be used for RNA extraction and subsequent phylogenomic studies. We sequenced 27 transcriptomes, 21 from silica gel-dried (SD-samples) and six from liquid nitrogen-preserved (LN-samples) leaf tissues, and downloaded 27 additional transcriptomes from GenBank. Our results showed that although the LN-samples produced slightly better reads than the SD-samples, there were no significant differences in RNA quality and quantity, assembled contig lengths and numbers, and BUSCO comparisons between two treatments. Using this data, we conducted phylogenomic analyses, including concatenated- and coalescent-based phylogenetic reconstruction, molecular dating, coalescent simulation, phylogenetic network estimation, and whole genome duplication (WGD) inference. The resulting phylogeny was consistent with previous studies with higher resolution and statistical support. The 11 core Ranunculaceae tribes grouped into two chromosome type clades (T- and R-types), with high support. Discordance among gene trees is likely due to hybridization and introgression, ancient genetic polymorphism and incomplete lineage sorting. Our results strongly support one ancient hybridization event within the R-type clade and three WGD events in Ranunculales. Evolution of the three Ranunculaceae chromosome types is likely not directly related to WGD events. By clearly resolving the Ranunculaceae phylogeny, we demonstrated that SD-samples can be used for RNA-seq and phylotranscriptomic studies of angiosperms.

evolutionary biology↗

bioRxiv2017

Defining a landscape of molecular phenotypes using a simple single sample scoring method

BackgroundGene set scoring provides a useful approach for quantifying concordance between sample transcriptomes and selected molecular signatures. Most methods use information from all samples to score an individual sample, leading to unstable scores in small data sets and introducing biases from sample composition across a data set (e.g. varying numbers of samples for different cancer subtypes). To address these issues we have developed a truly single sample scoring method, and associated R/Bioconductor package singscore.\n\nResultsWe have developed a rank-based single sample scoring method, implemented as a Bioconductor package. We use multiple cancer data sets to compare it against widely-used scoring methods, including GSVA, z-scores, PLAGE, and ssGSEA. Our approach does not depend upon background samples and thus the scores are stable regardless of the composition and number of samples in the gene expression data set. In contrast, scores obtained by GSVA, z-score, PLAGE and ssGSEA can be unstable when less data are available (nsamples < 25). We show that the computational time for singscore is faster than current implementations of GSVA and ssGSEA, and is comparable with that of z-score and PLAGE. The singscore package also produces visualisations and interactive plots that enable exploration of molecular phenotypes.\n\nConclusionsThe single sample scoring method described here is independent of sample composition in gene expression data and thus it provides stable scores that are less likely to be influenced by unwanted variation across samples. These scores can be used for dimensional reduction of transcriptomic data and the phenotypic landscapes obtained by scoring samples against multiple molecular signatures may provide insights for sample stratification.

bioinformatics↗