Biology subjects

An, J.-Y.

Publications and source records attributed to An, J.-Y..

13 recordsLinked to original sources

bioRxiv2026

AnnFlux: object-conditioned neural stochastic differential equations for single-cell perturbation dynamics

Single-cell perturbation profiling measures responses to genetic and chemical interventions, yet most models learn a static map, ignoring how populations move over time and how perturbations combine. AnnFlux, an object-conditioned stochastic differential equation, learns a drift field in latent cell-state space. Conditioning on the perturbing object makes the field queryable one object at a time, yielding per-object drifts comparable across genes and drugs. By learning a drift field tailored to each perturbation context, it interpolates a held-out timepoint in an epithelial-mesenchymal transition time course and predicts unseen perturbations. Beyond point estimates, AnnFlux improves distributional fidelity and predicts responses to held-out perturbation combinations. An IFN-response signature predicted by AnnFlux was associated with TLS proximity in an independent pan-cancer spatial atlas. This framework maps perturbation-driven cell-state evolution as continuous trajectories and represents unseen perturbations using prior-knowledge embeddings.

bioinformatics↗

bioRxiv2026

AI-driven framework modeling perturbation in brain organoids reveals candidate genes for autism

Autism gene discovery is constrained by the rarity and heterogeneity of damaging variants, requiring large cohorts to identify susceptibility genes. Neural organoids and single-cell foundation models enable perturbation modeling in neurodevelopmental contexts. Here, we show that perturbation-informed foundation modeling of neural organoids can provide functional context for prioritizing candidate genes with genomic and clinical support. We constructed a 3.6-million-cell organoid atlas and trained models to predict genome-wide perturbation responses. Benchmarking 17 models identified a telencephalic neuron-specific model best preserving autism-relevant perturbation structure. Genome-wide profiling revealed two clusters associated with mid-fetal synaptic neuronal processes and early radial glia ubiquitin signaling. These clusters were supported by damaging-variant enrichment and clinical phenotypes across 89,916 family-based samples. Logistic-regression prioritization identified 343 candidates, including 167 in the key clusters, with convergence across TADA signals and recurrent evidence for NBEA and KLHDC10. This framework integrates predicted perturbation effects with genomic evidence to support autism candidate prioritization.

bioinformatics↗

bioRxiv2026

megaMine: a scalable, rule-based framework for mining gene-cancer-drug evidence from biomedical literature

The rapid expansion of the oncology literature has outpaced manual curation of clinically relevant gene-cancer-drug associations and oncogenic driver evidence. Existing automated approaches often lack transparency or are difficult to scale across heterogeneous data sources. To address this gap, we developed megaMine, a transparent, rule-based, and context-aware literature-mining framework that integrates therapeutic and driver evidence from PubMed, PubTator, and Europe PMC by combining entity recognition, hierarchical heuristics, and contextual labeling. In therapy mode, megaMine was applied to approximately 100,000 oncology articles published between 2015 and 2025, yielding more than 23,000 structured sentence-level evidence records, with standardized annotations for drug response, resistance, and study context. Internal evaluation of context labels showed strong separability between efficacy and non-efficacy evidence using ridge logistic regression (AUROC = 0.915; AUPRC = 0.941). Benchmarking against NCI/OncoKB-supported drug-cancer associations showed that curated clinical associations had higher megaMine composite evidence scores than unlabeled comparison pairs [median (IQR): 25.6 (9.07-72.5) vs. 3.61 (1.69-8.69); Wilcoxon rank-sum test, P < 2.2 x 10^-16]. In driver mode, megaMine retrieved mutation- and biomarker-related evidence from an ERBB-focused gastric cancer query, generating 750 evidence rows from 200 PMIDs. These results demonstrate that deterministic and interpretable approaches can support scalable evidence extraction for downstream applications such as knowledge graph construction and literature-based evidence synthesis.

bioinformatics↗

bioRxiv2025

AGO1 in neural progenitor cells orchestrates brain development and sociability via LIN28A-REELIN axis

AGO1, an essential RNA binding protein (RBP) in RNA interference, is associated with autism spectrum disorder (ASD). However, the precise functions of AGO1 in brain development and related disorders remain largely unexplored. Here, we report the critical roles of AGO1 in neural progenitor cells (NPCs) in promoting sociability and shaping the structure of the developing brain. Ago1 knockout (KO) in the mouse brain leads to hyposociability, a characteristic symptom of ASD. In human forebrain organoids, AGO1 KO disrupts the formation of ventricle-like structures and delays cortical layer development. We discovered that AGO1 KO results in a loss of polarity in NPCs, which subsequently reduces neuronal development. These AGO1 KO phenotypes originate from the failure to suppress LIN28A in NPCs. AGO1 is primarily localized in the nucleus within NPCs and binds to the LIN28A promoter region, thereby inhibiting LIN28A transcription. We found that increased LIN28A in AGO1 KO NPCs reduces REELIN expression, a key regulator of brain development, by binding directly to the REELIN mRNA. The aberrant polarity phenotype in AGO1 KO NPCs was successfully ameliorated by either LIN28A knockdown or recombinant REELIN treatment. Collectively, our findings elucidate the intricate molecular mechanisms orchestrated by nuclear AGO1 in NPCs and underscores its pivotal roles in brain development, providing significant insights into ASD.

neuroscience↗

bioRxiv2025

Characterization of Human Anterior Neural Organoids as a Model for Investigating Cohen Syndrome

Neural organoids display three-dimensional (3D) structures that resemble in vivo neural architectures. Previously, we developed a novel two-dimensional (2D) neural induction-based protocol for culturing spinal cord organoids, enabling size control and recapitulating neural tube morphogenesis. In this study, we evaluated the application of this concept to induce the anterior regions of the brain and generate human anterior neural organoids (hANOs). By inducing neuroepithelial (NE) cells in 2D and re-aggregating them led to tube-forming morphogenesis similar to that posterior spinal cord induction. The transcriptome profiles of these hANOs resembled the frontal cortex of 20 weeks post-conception (PCW) human embryos. Using this hANOs protocol, we investigated microcephaly phenotypes associated with Cohen syndrome (CS), caused by biallelic loss-of-function variants in VPS13B gene. Deleting VPS13B in human pluripotent stem cells resulted in Golgi dispersion and growth retardation onset in mutant hANOs, akin to CS patients with postnatal microcephaly. This delay is partly linked to reduced neuronal growth. Additionally, mature CS organoids showed enhanced hyper-excitability associated with an excitatory/inhibitory imbalance. In conclusion, this protocol is suitable for studying microcephaly phenotypes from human genetic mutations due to its simplicity and scalability.

neuroscience↗

bioRxiv2024

An integrative single-cell atlas to explore the cellular and temporal specificity of neurological disorder genes during human brain development

Single-cell technologies have enhanced comprehensive knowledge regarding the human brain by facilitating an extensive transcriptomic census across diverse brain regions. Nevertheless, understanding the cellular and temporal specificity of neurological disorders remains ambiguous due to the developmental variations. To address this gap, we illustrated the dynamics of disorder risk gene expressions under development by integrating multiple single-cell RNA sequencing datasets. We constructed a comprehensive single-cell atlas of developing human brains, encompassing 393,060 single cells across diverse developmental stages. Temporal analysis revealed the distinct expression patterns of disorder risk genes, including autism, highlighting their temporal regulation in different neuronal and glial lineages. We identified distinct neuronal lineages diverged across developmental stages, each exhibiting temporal-specific expression patterns of disorder genes. Lineages of non-neuronal cells determined by molecular profiles also showed temporal-specific expressions, indicating a link between cellular maturation and the risk of disorder. Furthermore, we explored the regulatory mechanisms involved in early brain development, revealing enriched patterns of fetal cell types for neuronal disorders, indicative of the prenatal stages influence on disease determination. Our findings facilitate unbiased comparisons of cell type-disorder associations and provide insight into dynamic alterations in risk genes during development, paving the way for a deeper understanding of neurological disorders.

neuroscience↗

bioRxiv2024

Assessing Long-Term Stored Tissues for Multi-Omics Data Quality and Proteogenomics Suitability

As research into the complexities of cancer biology deepens, the integration of multi-omics analyses has emerged as a powerful approach to unravel the complex molecular basis of cancers. However, challenges related to sample availability, including size, collection procedures, and storage duration, hinder the broad application of this methodology. Despite these limitations, there is a growing interest in exploring the potential of archived samples to expand the scope of multi-omics research. Our study aims to investigate the impact of storage duration on the measurment in genomic, transcriptomic, and proteomic profiles of archived samples, demonstrating their viability for advancing our understanding of cancer biology. To comprehensively address these trends and limitations, we systematically examined archived samples collected over a decade, focusing on their transcriptomic, proteomic, and phosphoproteomic attributes. Analysis revealed intricate patterns and dynamic shifts, especially in long-term transcriptomic data, with observed declines in read counts related to protein coding and gene coverage. However, these changes did not compromise the fundamental gene expression landscape. Proteomic result also demonstrated that storage period did not significantly influence proteomic measurement. Comparisons of housekeeping gene (HKG) and housekeeping protein (HKP) expressions unveiled consistent transcriptomic levels across samples, while distinctive proteomic disparities between tumor and normal tissues. In conclusion, the challenges posed by limited sample availability in multi-omics studies can be partially overcome through the strategic integration of archived samples. While technical shifts were evident in certain aspects of transcriptomic data, core gene expression patterns remained robust, and the functionality of essential transcription factors (TFs) and kinases remained unaffected. These findings underscore the potential of archived samples as valuable resources for multi-omics research, providing a broader landscape for investigating cancer biology and paving the way for more comprehensive insights into this intricate field.

cancer biology↗

bioRxiv2021

A Cre-dependent massively parallel reporter assay allows for cell-type specific assessment of the functional effects of genetic variants in vivo

The function of regulatory elements is highly dependent on the cellular context, and thus for understanding the function of elements associated with psychiatric diseases these would ideally be studied in neurons in a living brain. Massively Parallel Reporter Assays (MPRAs) are molecular genetic tools that enable functional screening of hundreds of predefined sequences in a single experiment. These assays have not yet been adapted to query specific cell types in vivo in a complex tissue like the mouse brain. Here, using a test-case 3'UTR MPRA library with genomic elements containing variants from ASD patients, we developed a method to achieve reproducible measurements of element effects in vivo in a cell type-specific manner, using excitatory cortical neurons and striatal medium spiny neurons as test cases. This new technique should enable robust, functional annotation of genetic elements in the cellular contexts most relevant to psychiatric disease.

genetics↗

bioRxiv2020

Developmental dynamics of voltage-gated sodium channel isoform expression in the human and mouse neocortex

ObjectiveGenetic variants in the voltage-gated sodium channels SCN1A, SCN2A, SCN3A, and SCN8A are leading causes of epilepsy, developmental delay, and autism spectrum disorder. The mRNA splicing patterns of all four genes vary across development in the rodent brain, including mutually exclusive copies of the fifth protein-coding exon detected in the neonate (5N) and adult (5A). A second pair of mutually exclusive exons is reported in SCN8A only (18N and 18A). We aimed to quantify the expression of individual exons in the developing human neocortex. MethodsRNA-seq data from 176 human dorsolateral prefrontal cortex samples across development were analyzed to estimate exon-level expression. Developmental changes in exon utilization were validated by assessing intron splicing. Exon expression was also estimated in RNA-seq data from 58 developing mouse neocortical samples. ResultsIn the mature human neocortex, exon 5A is consistently expressed at least 4-fold higher than exon 5N in all four genes. For SCN2A, SCN3A, and SCN8A a synchronized 5N/5A transition occurs between 24 post-conceptual weeks (2nd trimester) and six years of age. In mice, the equivalent 5N/5A transition begins at or before embryonic day 15.5. In SCN8A, over 90% of transcripts in the mature human cortex include exon 18A. Early in fetal development, most transcripts include 18N or skip both 18N and 18A, with a transition to 18A inclusion occurring from 13 post-conceptual weeks to 6 months of age. No other protein-coding exons showed comparably dynamic developmental trajectories. SignificanceSplice isoforms, which alter the biophysical properties of the encoded channels, may account for some of the observed phenotypic differences across development and between specific variants. Manipulation of the proportion of splicing isoforms at appropriate stages of development may act as a therapeutic strategy for specific mutations or even epilepsy in general.

genetics↗

bioRxiv2020

Systematic evaluation of genome sequencing as a first-tier diagnostic test for prenatal and pediatric disorders

Current clinical guidelines recommend three genetic tests for the assessment of fetal structural anomalies: karyotype to detect microscopically-visible balanced and unbalanced chromosomal rearrangements, chromosomal microarray (CMA) to detect sub-microscopic copy number variants (CNVs), and exome sequencing (ES) to identify individual nucleotide changes in coding sequence. Advances in genome sequencing (GS) analysis suggest that it is poised to displace the sequential application of all three conventional tests to become a single diagnostic approach for the assessment of fetal structural anomalies. However, systematic benchmarking is required to assure that GS can capture the full mutational spectrum associated with fetal structural anomalies and to accurately quantify the added diagnostic yield of GS. We applied a novel GS analytic framework that included the discovery, filtration, and interpretation of nine classes of genomic variation to 7,195 individuals. We assessed the sensitivity of GS to detect diagnostic variants (pathogenic or likely pathogenic) from three standard-of-care tests using 1,612 autism spectrum disorder quartet families (ASD; n=6,448) with matched GS, ES, and CMA data, and validated these findings in 46 fetuses with a clinically reportable variant originally identified by karyotype, CMA, or ES. We then assessed the added diagnostic yield of GS in 249 trios (n=747) comprising a fetus with a structural anomaly detected by ultrasound and two unaffected parents that were pre-screened with a combination of all three standard-of-care tests. Across both cohorts, our GS analytic framework identified 98.2% of all diagnostic variants detected by standard-of-care tests, including 100% of those originally detected by CMA (n=88) and ES (n=61), as well as 78.6% (n=11/14) of the chromosomal rearrangements identified by karyotype. The diagnostic yield from GS was 7.8% across all 1,612 ASD probands, almost two-fold more than CMA (4.4%) and three-fold more than ES (3.0%). We also demonstrated that the yield of ES can approach that of GS when CNVs are captured with high sensitivity from exome data (7.4% vs. 7.8%, respectively). In 249 pre-screened fetuses with structural anomalies, GS provided an additional diagnostic yield of 0.4% beyond the combination of all three tests (karyotype, CMA, and ES). Applying our benchmarking results to existing data indicates that GS can achieve an overall diagnostic yield of 46.1% in unselected fetuses with fetal structural anomalies, providing an estimated 17.2% increase in diagnostic yield over karyotype, 14.1% over CMA, and 36.1% over ES when sequence variants are assessed, and 4.1% when CNVs are also identified from exome data. In this study we demonstrate that GS is sensitive to the detection of almost all pathogenic variation captured by karyotype, CMA, and ES, provides a superior diagnostic yield than any individual test by a wide margin, and contributes a modest increase in diagnostic yield beyond the combination of all three tests. We also outline several strategies to aid the interpretation of GS variants that are cryptic to conventional technologies, which we anticipate will be increasingly encountered as comprehensive variant identification from GS is performed. Taken together, these data suggest GS warrants consideration as a first-tier diagnostic approach for fetal structural anomalies.

genomics↗

bioRxiv2018

Novel genes for autism implicate both excitatory and inhibitory cell lineages in risk

We present the largest exome sequencing study to date focused on rare variation in autism spectrum disorder (ASD) (n=35,584). Integrating de novo and case-control variation with an enhanced Bayesian framework incorporating evolutionary constraint against mutation, we implicate 99 genes in ASD risk at a false discovery rate (FDR) [≤] 0.1. Of these 99 risk genes, 46 show higher frequencies of disruptive de novo variants in individuals ascertained for severe neurodevelopmental delay, while 50 show higher frequencies in individuals ascertained for ASD, and comparing ASD cases with disruptive mutations in the two groups shows differences in phenotypic presentation. Expressed early in brain development, most of the risk genes have roles in neuronal communication or regulation of gene expression, and 12 fall within recurrent copy number variant loci. In human cortex single-cell gene expression data, expression of the 99 risk genes is also enriched in both excitatory and inhibitory neuronal lineages, implying that disruption of these genes alters the development of both neuron types. Together, these insights broaden our understanding of the neurobiology of ASD.

genetics↗

bioRxiv2017

Whole Genome Sequencing in Psychiatric Disorders: the WGSPD Consortium

As technology advances, whole genome sequencing (WGS) is likely to supersede other genotyping technologies. The rate of this change depends on its relative cost and utility. Variants identified uniquely through WGS may reveal novel biological pathways underlying complex disorders and provide high-resolution insight into when, where, and in which cell type these pathways are affected. Alternatively, cheaper and less computationally intensive approaches may yield equivalent insights. Understanding the role of rare variants in the noncoding gene-regulating genome, through pilot WGS projects, will be critical to determine which of these two extremes best represents reality. With large cohorts, well-defined risk loci, and a compelling need to understand the underlying biology, psychiatric disorders have a role to play in this preliminary WGS assessment. The WGSPD consortium will integrate data for 18,000 individuals with psychiatric disorders, beginning with autism spectrum disorder, schizophrenia, bipolar disorder, and major depressive disorder, along with over 150,000 controls.

genomics↗

bioRxiv2017

Limited contribution of rare, noncoding variation to autism spectrum disorder from sequencing of 2,076 genomes in quartet families

Genomic studies to date in autism spectrum disorder (ASD) have largely focused on newly arising mutations that disrupt protein coding sequence and strongly influence risk. We evaluate the contribution of noncoding regulatory variation across the size and frequency spectrum through whole genome sequencing of 519 ASD cases, their unaffected sibling controls, and parents. Cases carry a small excess of de novo (1.02-fold) noncoding variants, which is not significant after correcting for paternal age. Assessing 51,801 regulatory classes, no category is significantly associated with ASD after correction for multiple testing. The strongest signals are observed in coding regions, including structural variation not detected by previous technologies and missense variation. While rare noncoding variation likely contributes to risk in neurodevelopmental disorders, no category of variation has impact equivalent to loss-of-function mutations. Average effect sizes are likely to be smaller than that for coding variation, requiring substantially larger samples to quantify this risk.

genomics↗