Biology subjects

Shim, H.

Publications and source records attributed to Shim, H..

At least 19 recordsLinked to original sources

bioRxiv2026

Stress granule fusion is a mitochondria-coordinated process for stress adaptation

Stress granules are cytoplasmic membraneless organelles assembled during stress to maintain cellular homeostasis. Although fusion is a hallmark of liquid-like behavior of these condensates, whether this process carries functional significance beyond its physical coalescence remains unclear. Here, we show that stress granule fusion is facilitated by mitochondrial dynamics and membrane potential. Intact mitochondria actively associate with stress granules, facilitating fusion and maturation. In contrast, loss of mitochondrial membrane potential, along with disrupted mitochondrial structure or motility, weakens these interactions and reduces fusion frequency. We find that impaired fusion leads to the accumulation of immature granules that retain fewer sequestered components, which correlates with premature cell death. Remarkably, restoring mitochondrial membrane potential rescues granule fusion and enlargement, and is accompanied by increased cell viability and a corresponding increase in granule-associated apoptotic factors. These results demonstrate that stress granule fusion is actively coordinated by mitochondria rather than driven solely by passive coalescence, reshaping how condensate dynamics are understood to integrate with organelle function during cellular stress response.

cell biology↗

bioRxiv2026

Temporal dynamics of microbiome communities within urban compost piles undergoing the heat process

Urban composting supports soil health but also intersects with food safety, where compost is produced near farms and communities. Here, we profiled temporal microbiome dynamics across a 6-week heat compost cycle from the urban compost piles using paired physicochemical panels and long-read metagenomics. Nutrient composition and pH shifted with compost age, coinciding with stage-structured microbial succession, including temperature-linked turnover of compost communities from mesophilic to thermotolerant taxa. Bacterial profiles included the presence of antimicrobial resistance genes and foodborne-associated genera early in the cycle, with reduced representation during the thermophilic phase. Analysis of previously unclassified long reads reveals an extensive repertoire of putative bacteriophages, including several complete genomes and candidates linked to foodborne bacteria, and their abundance is coupled to the host abundance. Together, these results support thermophilic composting as a key mitigation step for microbiological hazards in urban-adjacent systems and identify compost piles as a promising reservoir for discovering candidate lytic phages for downstream isolation and host-range testing.

microbiology↗

bioRxiv2025

Denoising image-based spatial transcriptomics data with DenoIST

Image-based spatial transcriptomics (IST) technologies provide unprecedented resolution of gene expression in tissue sections, but suffer from contamination of cells gene expression profiles due to imperfect cell segmentation. We present DenoIST (Denoising Image-based Spatial Transcriptomics data), a new computational tool that accurately identifies and removes contaminating transcripts from IST datasets. DenoIST models the observed transcript counts using a Poisson mixture model that explicitly accounts for local neighbourhood contamination. Applied to multiple real IST datasets of varying cell densities, DenoIST restores gene expression specificity and clarifies local biological structures by identifying and filtering transcripts spilt over from neighbouring cells. The denoised data enable more consistent and interpretable cell type annotation by minimising conflicting gene expression profiles, reducing the prevalence of hybrid or ambiguous cell types, and enhancing the contrast between distinct functional compartments. Overall, we demonstrate that DenoIST can be integrated to existing IST analysis workflows to improve biological interpretability and robustness of IST data.

bioinformatics↗

bioRxiv2025

Targeting multidrug-resistant bacteria with genetic-information-free protein-only phages

Bacteriophages offer advantages over small-molecule antibiotics, including host specificity and general compatibility with the human phagenome. However, their evolvability as replicating biological entities introduces therapeutic unpredictability and risks of phage-bacteria co-evolution. Here, we retain the targeting benefits of phages while avoiding genetic replication by engineering genetic-information-free, protein-only phages (POPs). These genome-free particles self-assemble in a cell-free protein synthesis system from modular, de novo gene fragments encoding only structural and antimicrobial proteins. Using Enterobacteria phage T7 and its susceptible bacterial host as a model, we test the hypothesis that POPs stochastically encapsulate small antimicrobial proteins during self-assembly and deliver them into bacteria during adsorption via an ejectome-mediated injection mechanism. A computational survey of T7 small proteins revealed early and mid-genome enrichments of hypothetical proteins and capsid volume sufficient to accommodate multiple small proteins in the absence of the [~]40-kb genome. In time-series antimicrobial susceptibility assays (48-72 h), POPs produced initial growth inhibition comparable to wild-type T7 at the highest doses with a linear dose-effect relationship and a minimum inhibitory concentration-like threshold. These results establish the feasibility of genetic-information-free POPs as protein-based antimicrobials that couple phage receptor specificity with minimal biosafety risks, supporting the development of more stable and predictable phage-inspired therapeutics.

pharmacology and toxicology↗

bioRxiv2025

SLAB: Simultaneous Labeling And Binding affinity prediction for protein-ligand structures

Machine learning models are often used as scoring functions to predict the binding affinity of a protein-ligand complex. These models are trained with limited amounts of data with experimentally measured binding affinity values. A large number of compounds are labeled inactive throughs single-concentration screens without measuring binding affinities. These inactive compounds, along with the active ones, can be used to train binary classification models, while regression models are trained using compounds with binding affinities only. However, the classification and regression tasks are often handled separately, without sharing the learned feature representations. In this paper, we propose a novel model architecture that jointly performs regression and classification objectives, aiming to maximize data utilization and improve predictive performance by leveraging two complementary tasks. In our setup, the regression yields the binding affinity, whereas the classification task yields the label as active or inactive. We demonstrate our method using PDBbind, the standard 3D structure database, as well as a dataset of flavivirus protease compounds with binding affinity data. Our experiments show that the new joint training strategy improves the accuracy of the model, increasing applicability in various practical drug screening scenarios.

pharmacology and toxicology↗

bioRxiv2024

Going beyond cell clustering and feature aggregation: Is there single cell level information in single-cell ATAC-seq data?

1Single-cell Assay for Transposase Accessible Chromatin with sequencing (scATAC-seq) has become a widely used method for investigating chromatin accessibility at single-cell resolution. However, the resulting data is highly sparse with most data entries being zeros. As such, currently available computational methods for scATAC-seq feature a range of transformation procedures to extract meaningful information from the sparse data. Most notably, these transformations can be categorized into: 1) feature aggregation with known biological associations, 2) pseudo-bulking cells of similar biology, and 3) binarisation of count data. These strategies beg the question of whether or not scATAC-seq data actually has usable single-cell and single-region information as intended from the assay. If we can go beyond aggregated features and pooled cells, it opens up the possibility of more complex statistical tasks that require that degree of granularity. To reach the finest possible resolution of single-cell, single-region information there are inevitably many computational challenges to overcome. Here, we review the major data analysis challenges lying between raw data readout and biological discovery, and discuss the limitations of current data analysis approaches. Lastly, we conclude that chromatin accessibility profiling at true single-cell resolution is not yet achieved with current technology, but that it may be achieved with promising developments in optimising the efficiency of scATAC-seq assays.

bioinformatics↗

bioRxiv2024

Nanopore sequencing from protozoa to phages: decoding biological information on a string of biochemical molecules into human-readable signals

Biological information is encoded in a sequence of biochemical molecules such as nucleic acids and amino acids, and nanopore sequencing is a long-read sequencing technology capable of directly decoding these molecules into human-readable signals. The long reads from nanopore sequencing offer the advantage of obtaining contiguous information, which is particularly beneficial for decoding complex or repetitive regions in a genome. In this study, we investigated the efficacy of nanopore sequencing in decoding biological information from distinctive genomes in metagenomic samples, which pose significant challenges for traditional short-read sequencing technologies. Specifically, we sequenced blood and fecal samples from mice infected with Trypanosoma brucei, a unicellular protozoan known for its hypervariable and dynamic regions that help it evade host immunity. Such characteristics are also prevalent in other host-dependent parasites, such as bacteriophages. The taxonomic classification results showed a high proportion of nanopore reads identified as T. brucei in the infected blood samples, with no significant identification in the control blood samples and fecal samples. Furthermore, metagenomic de novo assembly of these nanopore reads yielded contigs that mapped to the reference genome of T. brucei in the infected blood samples with over 96% accuracy. This exploratory work demonstrates the potential of nanopore sequencing for the challenging task of classifying and assembling hypervariable and dynamic genomes from metagenomic samples.

bioengineering↗

bioRxiv2024

A pangenome analysis of ESKAPE bacteriophages: the underrepresentation may impact machine learning models

Bacteriophages are the most prevalent biological entities in the biosphere. However, limitations in both medical relevance and sequencing technologies have led to a systematic underestimation of the genetic diversity within phages. This underrepresentation not only creates a significant gap in our understanding of phage roles across diverse biosystems but also introduces biases in computational models reliant on these data for training and testing. In this study, we focused on publicly available genomes of bacteriophages infecting high-priority ESKAPE pathogens to show the extent and impact of this underrepresentation. First, we demonstrate a stark underrepresentation of ESKAPE phage genomes within the public genome and protein databases. Next, a pangenome analysis of these ESKAPE phages reveals extensive sharing of core genes among phages infecting the same host. Furthermore, genome analyses and clustering highlight close nucleotide-level relationships among the ESKAPE phages, raising concerns about the limited diversity within current public databases. Lastly, we uncover a scarcity of unique lytic phages and phage proteins with antimicrobial activities against ESKAPE pathogens. This comprehensive analysis of the ESKAPE phages underscores the severity of underrepresentation and its potential implications. This lack of diversity in phage genomes may restrict the resurgence of phage therapy and cause biased outcomes in data-driven computational models due to incomplete and unbalanced biological datasets.

bioinformatics↗

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

NAD: Noise-augmented direct sequencing of target nucleic acids by augmenting with noise and selective sampling

Next-generation sequencing necessitates a minimum quantity and concentration of DNA/RNA samples, typically achieved through amplification using the PCR technique. However, this amplification step introduces several drawbacks to biological insights, including PCR bias and the loss of epigenetic information. The advent of long-read sequencing technologies facilitates direct sequencing, with the primary constraint being the limited amount of DNA/RNA present in biological samples. Here, we present a novel method called Noise-Augmented Direct (NAD) sequencing that enables the direct sequencing of target DNA even when it falls below the minimum quantity and concentration required for long-read sequencing by augmenting with noise DNA and adaptive sampling. Adaptive sampling is an emerging technology of nanopore sequencing, allowing the enhanced sequencing of target DNA by selectively depleting noise DNA. In this study, we use the DNA standard of the Lambda phage genome as the noise DNA to augment samples containing low amounts of bacterial genomes (1 ng to 300 ng). The results with cost-effective flow cells indicate that NAD sequencing successfully detects the target DNA with an input quantity as low as 1 ng, and the bacterial genome of Salmonella enterica can be assembled to 30% completion at an accuracy of 98% with an input quantity of 3 ng. With high throughput flow cells, the bacterial genome of Pseudonomas aeruginas was assembled to near completion (99.9%) at an accuracy of 99.97% with an input quantity of 300 ng. This proof-of-concept study demonstrates the potential of NAD sequencing in enhancing the robustness of long-read sequencing with small input DNA/RNA samples with noise augmentation and adaptive sampling.

bioengineering↗

bioRxiv2023

Sequencing, fast and slow: profiling microbiomes in human samples with nanopore sequencing

Rapid and accurate pathogen identification is crucial in effectively combating infectious diseases. However, the current diagnostic tools for bacterial infections predominantly rely on century-old culture-based methods. Furthermore, recent research highlights the significance of host-microbe interactions within the host microbiota in influencing the outcome of infection episodes. As our understanding of science and medicine continues to advance, there is a pressing need for innovative diagnostic methods that can identify pathogens and also rapidly and accurately profile the microbiome landscape in human samples. In clinical settings, such diagnostic tools will become a powerful predictive instrument in directing the diagnosis and prognosis of infectious diseases by providing comprehensive insights into the patients microbiota. Here, we explore the potential of long-read sequencing in profiling the microbiome landscape from various human samples in terms of speed and accuracy. Using nanopore sequencers, we generate native DNA sequences from saliva and stool samples rapidly, from which each long-read is basecalled in real-time to provide downstream analyses such as taxonomic classification and antimicrobial resistance through the built-in software (< 12 hours). Subsequently, we utilize the nanopore sequence data for in-depth analysis of each microbial species in terms of host-microbe interaction types and deep learning-based classification of unidentified reads. We find that the nanopore sequence data encompass complex information regarding the microbiome composition of the host and its microbial communities, and also shed light on the unexplored human mobilome including bacteriophages. In this study, we use two different systems of long-read sequencing to give insights into human microbiome samples in the slow and fast modes, which raises additional inquiries regarding the precision of this novel technology and the feasibility of extracting native DNA sequences from other human microbiomes.

microbiology↗

bioRxiv2023

CRISPR-Cas-Docker: Web-based in silico docking and machine learning-based classification of crRNAs with Cas proteins

MotivationCRISPR-Cas-Docker is a web server for in silico docking experiments with CRISPR RNAs (crRNAs) and Cas proteins. This web server aims at providing experimentalists with the optimal crRNA-Cas pair predicted computationally when prokaryotic genomes have multiple CRISPR arrays and Cas systems, as frequently observed in metagenomic data. CRISPR-Cas-Docker provides two methods to predict the optimal Cas protein given a particular crRN sequence: a structure-based method (in silico docking) and a sequence-based method (machine learning classification). For the structure-based method, users can either provide experimentally determined 3D structures of these macromolecules or use an integrated pipeline to generate 3D-predicted structures for in silico docking experiments. ResultsCRISPR-Cas-Docker is an optimized and integrated platform that provides users with 1) 3D-predicted crRNA structures and AlphaFold-predicted Cas protein structures, 2) the top-10 docking models for a particular crRNA-Cas protein pair, and 3) machine learning-based classification of crRNA into its Cas system type. Availability and implementationCRISPR-Cas-Docker is available as an open-source tool under the GNU General Public License v3.0 on GitHub. It is also available as a web server.

bioinformatics↗

bioRxiv2022

Identification of cell barcodes from long-read single-cell RNA-seq with BLAZE

Single-cell RNA sequencing (scRNA-seq) has revolutionised our ability to profile gene expression. However, short-read (SR) scRNAseq methodologies such as 10x are restricted to sequencing the 3 or 5 ends of transcripts, providing accurate gene expression but little information on the RNA isoforms expressed in each cell. Newly developed long-read (LR) scRNA-seq enables the quantification of RNA isoforms in individual cells but LR scRNA-seq using the Oxford Nanopore platform has largely relied upon matched short-read data to identify cell barcodes and allow single cell analysis. Here we introduce BLAZE (Barcode identification from long-reads for AnalyZing single-cell gene Expression), which accurately and efficiently identifies 10x cell barcodes using only nanopore LR scRNA-seq data. We compared BLAZE to existing tools, including cell barcodes identified from matched SR scRNA-seq, on differentiating stem cells and 5 cancer cell lines. BLAZE outperforms existing tools and provides a more accurate representation of the cells present in LR scRNA-seq than using matched short-reads. BLAZE provides accurate cell barcodes over a wide range of experimental read depths and sequencing accuracies, while other methodologies commonly identify false-positive barcodes and cell clusters, disrupting biological interpretation of LR scRNA-seq results. In conclusion, BLAZE eliminates the requirement for matched SR scRNA-seq to interpret LR scRNA-seq, simplifying procedures and decreasing costs while also improving LR scRNA-seq results. BLAZE is compatible with downstream tools accepting a cell barcode whitelist file and is available at https://github.com/shimlab/BLAZE.

bioinformatics↗

bioRxiv2022

Trade-off between conservation of biological variation and batch effect removal in deep generative modeling for single-cell transcriptomics

Single-cell RNA sequencing (scRNA-seq) technology has contributed significantly to diverse research areas in biology, from cancer to development. Since scRNA-seq data is high-dimensional, a common strategy is to learn low-dimensional latent representations better to understand overall structure in the data. In this work, we build upon scVI, a powerful deep generative model which can learn biologically meaningful latent representations, but which has limited explicit control of batch effects. Rather than prioritizing batch effect removal over conservation of biological variation, or vice versa, our goal is to provide a birds eye view of the trade-offs between these two conflicting objectives. Specifically, using the well established concept of Pareto front from economics and engineering, we seek to learn the entire trade-off curve between conservation of biological variation and removal of batch effects. A multi-objective optimisation technique known as Pareto multi-task learning (Pareto MTL) is used to obtain the Pareto front between conservation of biological variation and batch effect removal. Our results indicate Pareto MTL can obtain a better Pareto front than the naive scalarization approach typically encountered in the literature. In addition, we propose to measure batch effect by applying a neural-network based estimator called Mutual Information Neural Estimation (MINE) and show benefits over the more standard Maximum Mean Discrepancy (MMD) measure. The Pareto front between conservation of biological variation and batch effect removal is a valuable tool for researchers in computational biology. Our results demonstrate the efficacy of applying Pareto MTL to estimate the Pareto front in conjunction with applying MINE to measure the batch effect.

bioinformatics↗

bioRxiv2022

In silico optimization of RNA-protein interactions for CRISPR-Cas13-based antimicrobials

RNA-protein interactions are crucial for diverse biological processes. In prokaryotes, RNA-protein interactions enable adaptive immunity through CRISPR-Cas systems. These defense systems utilize CRISPR RNA (crRNA) templates acquired from past infections to destroy foreign genetic elements through crRNA-mediated nuclease activities of Cas proteins. Thanks to the programmability and specificity of CRISPR-Cas systems, CRISPR-based antimicrobials have the potential to be repurposed as new types of antibiotics. Unlike traditional antibiotics, these CRISPR-based antimicrobials can be designed to target specific bacteria and minimize detrimental effects on the human microbiome during antibacterial therapy. Here, we explore the potential of CRISPR-based antimicrobials by optimizing the RNA-protein interactions of crRNAs and Cas13 proteins. CRISPR-Cas13 systems are unique as they degrade specific foreign RNAs using the crRNA template, which leads to non-specific RNase activities and cell cycle arrest. We show that a high proportion of the Cas13 systems have no colocalized CRISPR arrays, and the lack of direct association between crRNAs and Cas proteins may result in suboptimal RNA-protein interactions in the current tools. Here, we investigate the RNA-protein interactions of the Cas13-based systems by curating the validation dataset of Cas13 protein and CRISPR repeat pairs that are experimentally validated to interact, and the candidate dataset of CRISPR repeats that reside on the same genome as the currently known Cas13 proteins. To find optimal CRISPR-Cas13 interactions, we first validate the 3-D structure prediction of crRNAs based on their experimental structures. Next, we test a number of RNA-protein interaction programs to optimize the in silico docking of crRNAs with the Cas13 proteins. From this optimized pipeline, we find a number of candidate crRNAs that have comparable or better in silico docking with the Cas13 proteins of the current tools. This study fully automatizes the in silico optimization of RNA-protein interactions as an efficient preliminary step for designing effective CRISPR-Cas13-based antimicrobials.

bioinformatics↗

bioRxiv2022

Topology of gene regulatory compartments relative to the nuclear matrix

Summary ParagraphInside the nucleus, there are special compartments with characterized functions, some of which are involved in gene expression1. These compartments include transcriptional condensates and nuclear speckles, which contain factors required for transcription and splicing, respectively. While the characteristics of these intranuclear compartments were extensively investigated, spatial relationship between them is yet unclear. Meanwhile, RNA-protein structural network named the nuclear matrix has been observed inside the nucleus and suggested as a framework for the spatial segregation of the intranuclear space2-6. However, concept of the nuclear matrix was not widely accepted due to its lack of in vivo evidence7. Here, we report visualization of the nuclear matrix and intranuclear compartments using super-resolution fluorescence microscopy. We found the nuclear matrix is dynamic in live cells and easily disrupted upon transcription inhibition. Remarkably, we observed an orderly layered distribution of transcriptional condensates and nuclear speckles relative to the nuclear matrix. We also observed a separation of chromosome territories from transcriptional condensates and nuclear speckles by the nuclear matrix. Based on our findings, we propose a topological model for the regulation of transcription across the nuclear matrix.

cell biology↗

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

Rethinking protein drug design with highly accurate structure prediction of anti-CRISPR proteins

Protein therapeutics play an important role in controlling the functions and activities of disease-causing proteins in modern medicine. Despite protein therapeutics having several advantages over traditional small-molecule therapeutics, further development has been hindered by drug complexity and delivery issues. However, recent progress in deep learning-based protein structure prediction approaches such as AlphaFold opens new opportunities to exploit the complexity of these macro-biomolecules for highly-specialised design to inhibit, regulate or even manipulate specific disease-causing proteins. Anti-CRISPR proteins are small proteins from bacteriophages that counter-defend against the prokaryotic adaptive immunity of CRISPR-Cas systems. They are unique examples of natural protein therapeutics that have been optimized by the host-parasite evolutionary arms race to inhibit a wide variety of host proteins. Here, we show that these Anti-CRISPR proteins display diverse inhibition mechanisms through accurate structural prediction and functional analysis. We find that these phage-derived proteins are extremely distinct in structure, some of which have no homologues in the current protein structure domain. Furthermore, we find a novel family of Anti-CRISPR proteins which are structurally homologous to the recently-discovered mechanism of manipulating host proteins through enzymatic activity, rather than through direct inference. Using highly accurate structure prediction, we present a wide variety of protein-manipulating strategies of anti-CRISPR proteins for future protein drug design.

bioinformatics↗