Biology subjects

Liu, Y. H.

Publications and source records attributed to Liu, Y. H..

11 recordsLinked to original sources

bioRxiv2026

CNVeil resolves haplotype-specific copy number and uncovers subclonal architecture hidden from total copy number profiling in single-cell cancergenomes

Single-cell DNA sequencing (scDNA-seq) resolves copy number variation (CNV) at single-cell resolution, revealing tumor heterogeneity and subclonal structure. Most existing methods, however, infer only total copy number. Haplotype-resolved copy number, which captures allelic imbalance and clonal evolution, remains far less developed, largely because low coverage, allelic dropout, and technical noise in scDNA-seq make phased allelic inference substantially harder than total copy number estimation. We present CNVeil, a haplotype-aware framework that infers total, allele-specific, and chromosome-scale haplotype-resolved copy number from scDNA-seq data. CNVeil first builds robust total copy number profiles through highly variable bin selection, hierarchical clustering, subclone-aware ploidy estimation, and cross-cell consensus segmentation. Using this profile as a stable scaffold, it infers allele-specific copy number with an expectation-maximization algorithm applied to heterozygous SNP allele counts, then reconstructs haplotype-specific copy number by enforcing coherent haplotype orientation across adjacent segments via dynamic programming. We benchmarked CNVeil against 12 state-of-the-art methods, including eight total copy number callers, two allele-specific callers, and two haplotype-resolved callers, across 20 simulated and real datasets spanning six experimental settings, including high-multiplexed single-nucleus sequencing, Acoustic Cell Tagmentation (ACT), and 10x Chromium. This constitutes the largest comparative evaluation of single-cell copy number inference methods to date. CNVeil consistently outperformed existing tools in segmentation accuracy, ploidy inference, subclone identification, and allele-specific copy number estimation. In a breast cancer multi-omics (wellDR-seq) cohort, CNVeil uncovered haplotype-specific subclonal diversification invisible to total copy number analysis alone and linked allele-specific copy number states to transcriptional variation. By transforming sparse single-cell allelic signals into chromosome-scale haplotype-resolved profiles, CNVeil closes a major methodological gap and provides a scalable framework for studying tumor evolution and functional genomic heterogeneity at single-cell resolution.

bioinformatics↗

bioRxiv2026

Benchmarking cell type annotation in spatial transcriptomics: resolving cellular hierarchies, biological fidelity, and dynamic cell states

Spatial transcriptomics enables the quantification of gene expression within its native tissue context, providing unprecedented insight into tissue architecture, cellular ecosystems, and local cell-cell interactions at regional and single-cell resolution. Accurate cell type annotation is a critical prerequisite for interpreting these data and is often the first and most essential step in downstream analysis. Despite rapid advances in computational methods, cell type annotation remains challenging and frequently requires extensive expert-driven manual curation based on marker-gene expression, spatial context, and prior biological knowledge. While early approaches relied primarily on transcriptional similarity, newer methods increasingly incorporate spatial information, histological features, and multimodal data to improve annotation accuracy. Nevertheless, reliable annotation remains difficult when biological interpretation requires fine-grained subtype resolution, particularly for platforms with limited gene panels, tissues undergoing dynamic cellular state transitions, and studies in which reference and query datasets differ substantially in biological context or technical modality. Here, we present a systematic benchmark of 20 state-of-the-art cell type annotation methods across four spatial transcriptomics datasets spanning diverse technologies, experimental conditions, cell numbers, and gene panel sizes. Importantly, all benchmark datasets contain expert-curated cell type labels, including wellresolved cell populations and subtype annotations, providing high-quality biological ground truth for evaluation. The benchmark encompasses both reference-based and reference-free methods representing a broad range of computational frameworks. Performance was assessed using conventional classification metrics, including accuracy and F1-based measures, together with structure-aware metrics that evaluate both cell-level annotation accuracy and preservation of higher-order biological organization. Across datasets, annotation performance varied substantially according to tissue context, reference-query similarity, and annotation granularity. Fine-grained subtype annotation and recovery of rare cell populations remained challenging for many methods, particularly in datasets capturing injury, repair, developmental, and regenerative processes characterized by continuous cellular state transitions. Notably, high classification accuracy did not necessarily correspond to preservation of global cellular relationships or biologically coherent downstream pathway and gene-set enrichment analyses. Overall, scANVI, Seurat, and TACCO consistently ranked among the top-performing methods, although their relative advantages were context dependent. Together, our results provide a comprehensive assessment of current annotation strategies for spatial transcriptomics and offer practical guidance for selecting methods that best align with specific biological questions, dataset characteristics, and analytical priorities.

bioinformatics↗

bioRxiv2025

Effect of PAPE Induced by Different Squat Loads on Jump Performance in Collegiate Female Basketball Players

BackgroundPost-activation performance enhancement (PAPE) can acutely augment explosive performance, yet the optimal prescription of squat load and recovery for female basketball players remains unclear. MethodsTwenty-eight collegiate womens basketball athletes were randomly allocated to 90%, 80%, or 70% of one-repetition maximum (1RM) back-squat groups, or a control group. Athletes completed 3x3 parallel back squats, after which countermovement jump (CMJ), single-leg and double-leg approach-jump heights were assessed at baseline and 4, 8, and 12 min. CMJ kinetics--peak power output, vertical ground-reaction force, and flight time--and lower-limb surface EMG were recorded concurrently. ResultsLoads [≥]80%1RM reliably elicited PAPE, with improvements demonstrating clear load-dependence and time specificity. Double-leg approach-jump height peaked at 8 min in the 90%1RM group (40.50 {+/-} 1.73 cm). Single-leg approach-jump height was maximized at 8 min in the 80%1RM group (40.00 {+/-} 0.82 cm), exceeding the control condition. Kinetic and EMG analyses indicated that 90%1RM produced a delayed rectus femoris activation peak (8 min) and a later CMJ power peak (12 min), whereas 80%1RM facilitated earlier gastrocnemius activation (4 min) with a stable output profile; performance benefits with 70%1RM were minimal. ConclusionsA moderately high load (80%1RM) paired with an 8-12-min recovery window strikes a favorable balance between neural potentiation and fatigue, offering a practical pre-competition activation strategy to enhance anaerobic power and jump performance in female basketball players.

physiology↗

bioRxiv2025

Enhancing variant detection in complex genomes: leveraging linked reads for robust SNP, Indel, and structural variant analysis

Accurate detection of genetic variants, including single nucleotide polymorphisms (SNPs), small insertions and deletions (INDELs), and structural variants (SVs), is essential for comprehensive genomic analysis. While short-read sequencing performs well for SNP and INDEL detection, it remains limited in resolving SVs, particularly in complex genomic regions, due to its short read length. Linked-read sequencing technologies, such as single-tube Long Fragment Read (stLFR), partially address this limitation by incorporating molecular barcodes to provide long-range information. In this study, we evaluate conventional paired-end linked reads (PE100_stLFR) and explore a conceptual extension: long single-end barcoded reads of 500 bp (SE500_stLFR) and 1000 bp (SE1000_stLFR). We developed_stLFR-sim, a Python-based simulator that reproduces the_stLFR workflow and enables realistic benchmarking. Using a high-quality T2T assembly of HG002, we generated multiple datasets across 12 sequencing configurations. SVs were called using Aquila_stLFR (v2) and benchmarked against the Genome in a Bottle (GIAB) HG002 SV truth set with Truvari. We show that simulated PE100_stLFR closely matches real data, validating the simulation framework. Increasing read length consistently improves SV detection accuracy, with SE1000_stLFR achieving the best performance and approaching long-read methods while outperforming short-read and pangenome-based approaches. Collectively, our results highlight the strong potential of long single-end barcoded reads for improving SV detection, and suggest that even modest increases in read length, when combined with barcode information, can provide a cost-effective and practical strategy for enhancing future sequencing technologies and SV discovery.

bioinformatics↗

bioRxiv2024

Spiking neural network models of sound localisation via a massively collaborative process

AbstractNeuroscientists are increasingly initiating large-scale collaborations which bring together tens to hundreds of researchers. However, while these projects represent a step-change in scale, they retain a traditional structure with centralised funding, participating laboratories and data sharing on publication. Inspired by an open-source project in pure mathematics, we set out to test the feasibility of an alternative structure by running a grassroots, massively collaborative project in computational neuroscience. To do so, we launched a public Git repository, with code for training spiking neural networks to solve a sound localisation task via surrogate gradient descent. We then invited anyone, anywhere to use this code as a springboard for exploring questions of interest to them, and encouraged participants to share their work both asynchro-nously through Git and synchronously at monthly online workshops. At a scientific level, our work investigated how a range of biologically-relevant parameters, from time delays to mem-brane time constants and levels of inhibition, could impact sound localisation in networks of spiking units. At a more macro-level, our project brought together 31 researchers from multiple countries, provided hands-on research experience to early career participants, and opportunities for supervision and teaching to later career participants. Looking ahead, our project provides a glimpse of what open, collaborative science could look like and provides a necessary, tentative step towards it.

neuroscience↗

bioRxiv2024

Transition to chaos separates learning regimes and relates to measure of consciousness in recurrent neural networks

AbstractRecurrent neural networks exhibit chaotic dynamics when the variance in their connection strengths exceed a critical value. Recent work indicates connection variance also modulates learning strategies; networks learn "rich" representations when initialized with low coupling and "lazier"solutions with larger variance. Using Watts-Strogatz networks of varying sparsity, structure, and hidden weight variance, we find that the critical coupling strength dividing chaotic from ordered dynamics also differentiates rich and lazy learning strategies. Training moves both stable and chaotic networks closer to the edge of chaos, with networks learning richer representations before the transition to chaos. In contrast, biologically realistic connectivity structures foster stability over a wide range of variances. The transition to chaos is also reflected in a measure that clinically discriminates levels of consciousness, the perturbational complexity index (PCIst). Networks with high values of PCIst exhibit stable dynamics and rich learning, suggesting a consciousness prior may promote rich learning. The results suggest a clear relationship between critical dynamics, learning regimes and complexity-based measures of consciousness.

neuroscience↗

bioRxiv2024

MaskGraphene: Advancing joint embedding, clustering, and batch correction for spatial transcriptomics using graph-based self-supervised learning

Recent advancements in spatial transcriptomics (ST) have underscored the importance of integrating data from multiple ST slices for joint analysis. A major challenge remains generating interpretable joint embeddings that preserve geometric information for downstream analyses. Here we introduce MaskGraphene, a graph neural network that combines self-supervised and self-contrastive training to integrate gene expression and spatial location into joint embeddings. By employing clusterwise alignment and a graph attention autoencoder with masked self-supervised and triplet loss optimizations, MaskGraphene effectively preserves geometric structures while achieving batch correction. In benchmarks against seven state-of-the-art methods, MaskGraphene consistently demonstrated superior alignment accuracy and geometric fidelity across diverse ST datasets. Its interpretable embeddings significantly enhanced downstream applications, including domain identification, spatial trajectory reconstruction, biomarker discovery, and the creation of topographical maps of brain slices. Notably, MaskGraphene successfully recovered layer-wise brain structures with near-perfect accuracy. MaskGraphene provides a powerful and versatile framework for advancing ST data integration and analysis, unlocking valuable biological insights.

bioinformatics↗

bioRxiv2024

CNVeil enables accurate and robust tumor subclone identification and copy number estimation from single-cell DNA sequencing data

Single-cell DNA sequencing (scDNA-seq) has significantly advanced cancer research by enabling precise detection of chromosomal aberrations, such as copy number variations (CNVs), at a single-cell level. These variations are crucial for understanding tumor progression and heterogeneity among tumor subclones. However, accurate CNV inference in scDNA-seq has been constrained by several factors, including low coverage, sequencing errors, and data variability. To address these challenges, we introduce CNVeil, a robust quantitative algorithm designed to accurately reveal CNV profiles while overcoming the inherent noise and bias in scDNA-seq data. CNVeil incorporates a unique bias correction method using normal cell profiles identified by a PCA-based Gini coefficient, effectively mitigating sequencing bias. Subsequently, a multi-level hierarchical clustering, based on selected highly variable bins, is employed to initially identify coarse subclones for robust ploidy estimation and further identify fine subclones for segmentation. To infer the CNV segmentation landscape, a novel change rate-based across-cell breakpoint identification approach is specifically designed to diminish the effects of low coverage and data variability on a per-cell basis. Finally, a consensus segmentation is utilized to further standardize read depth for the inference of the final CNV profile. In comprehensive benchmarking experiments, where we compared CNVeil with seven state-of-the-art CNV detection tools, CNVeil exhibited exceptional performance across a diverse set of simulated and real scDNA-seq data in cancer genomics. CNVeil excelled in subclone identification, segmentation, and CNV profiling. In light of these results, we anticipate that CNVeil will significantly contribute to single-cell CNV analysis, offering enhanced insights into chromosomal aberrations and genomic complexity.

bioinformatics↗

bioRxiv2022

Ultrastructure of synaptic connectivity within sub-regions of the SCN revealed by genetically encoded EM tag and SBEM

The suprachiasmatic nucleus (SCN) in the hypothalamus of the vertebrate brain is the central pacemaker regulating circadian rhythmicity throughout the body. The SCN receives photic information through melanopsin-expressing retinal ganglion cells (mRGC) to synchronize the body with environmental light cycles. Determining how these inputs fit into the network of synaptic connections on and between SCN neurons is key to impelling our understanding of the regulation of the circadian clock by light and unraveling the relevant local circuits within the SCN. To map these connections, we used a newly-developed Cre-dependant electron microscopy reporter, APEX2, to label mitochondria of mRGC axons, and serial blockface scanning electron microscopy to resolve the fine structure of mRGC in 3D volumes of the SCN. The maps thus created provide a first draft of the patterns of connectomic organization of SCN in the core and the shell, composed of different neuronal subtypes, and here shown to differ with regard to the patterning of their mRGC input as the shell receives denser mRGCs synaptic inputs compared to the core. This challenges the presently held view that photic information coming directly from the retina is mainly integrated by the core region of the SCN.

neuroscience↗

bioRxiv2020

A solution to temporal credit assignment using cell-type-specific modulatory signals

Animals learn and form memories by jointly adjusting the efficacy of their synapses. How they efficiently solve the underlying temporal credit assignment problem remains elusive. Here, we re-analyze the mathematical basis of gradient descent learning in recurrent spiking neural networks (RSNNs) in light of the recent single-cell transcriptomic evidence for cell-type-specific local neuropeptide signaling in the cortex. Our normative theory posits an important role for the notion of neuronal cell types and local diffusive communication by enabling biologically plausible and efficient weight update. While obeying fundamental biological constraints, including separating excitatory vs inhibitory cell types and observing connection sparsity, we trained RSNNs for temporal credit assignment tasks spanning seconds and observed that the inclusion of local modulatory signaling improved learning efficiency. Our learning rule puts forth a novel form of interaction between modulatory signals and synaptic transmission. Moreover, it suggests a computationally efficient learning method for bio-inspired artificial intelligence.

neuroscience↗

bioRxiv2020

Emergence of prefrontal neuron maturation properties by training recurrent neural networks in cognitive tasks

Working memory and response inhibition are functions that mature relatively late in life, after adolescence, paralleling the maturation of the prefrontal cortex. The link between behavioral and neural maturation is not obvious, however, making it challenging to understand how neural activity underlies the maturation of cognitive function. To gain insights into the nature of observed changes in prefrontal activity between adolescence and adulthood, we investigated the progressive changes in unit activity of Recurrent Neural Networks (RNNs) as they were trained to perform working memory and response inhibition tasks. These included increased delay period activity during working memory tasks, and increased activation in antisaccade tasks. These findings reveal universal properties underlying the neuronal computations behind cognitive tasks and explicate the nature of changes that occur as the result of developmental maturation.

neuroscience↗