Biology subjects

Oh, M.

Publications and source records attributed to Oh, M..

18 recordsLinked to original sources

bioRxiv2026

Food perception alone stimulates oviposition through serotonin signaling in Drosophila

Reproduction is strongly shaped by nutrient availability, but whether sensory information about food can regulate reproductive output before nutrients are consumed remains unclear. Here, using Drosophila melanogaster, we show that sensory perception of dietary yeast, a protein-rich food source widely used to stimulate fecundity in the fruit fly, is sufficient to rapidly promote egg laying. Specifically, we found that yeast odor exposure increased the rate of oviposition and the number of eggs laid within hours, independent of yeast intake, and that this response was mediated by olfactory signaling through the odorant receptor co-receptor Orco. We further identified serotonin receptor 2A (5-HT2A), but not 5-HT2B, as necessary for yeast odor-induced egg laying. We then asked whether yeast odor also elicits broader life-history changes and whether its reproductive and longevity effects depend on a shared mechanism. Consistent with previous observations, yeast odor also shortened lifespan. However, this lifespan effect was independent of 5-HT2A signaling, indicating that the reproductive and longevity responses to food perception are separable. Finally, our brain connectomic analyses revealed that yeast-responsive olfactory neurons preferentially influence oviposition descending neurons (OviDNs) compared with other olfactory neurons. OviDNs further connect with endocrine-related circuitry in the corpora allata (CA), and, most interestingly, the interneurons linking OviDNs to CA neurons are enriched for serotonergic and dopaminergic modulation. Taken together, our observations indicate that nutrient acquisition is not a prerequisite for animals to adjust their reproduction to the nutritional environment. Sensory information predicting food availability itself can elicit changes in neural state, likely through olfactory-reproductive circuits, that acutely alter reproductive physiology.

physiology↗

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

Asymmetry and Allostery: Insights into the Mechanism of DirectionalPeptide Translocation in AAA+ Unfoldases

Molecular motors convert stochastic ATP chemistry into directional motion. Modern nonequilibrium theories explain this behavior through kinetic asymmetry, yet the molecular mechanisms that generate kinetic asymmetry in AAA+ unfoldases remain unresolved. Cryo-electron microscopy has consistently revealed unfoldases take on an asymmetric helical staircase structure, inspiring the proposed processive hand-over-hand translocation mechanism. In contrast, biochemical and single-molecule studies have reported bursts, slips, and variable step sizes indicative of stochastic behavior. How these observations are reconciled and how directional motion emerges from stochastic ATP hydrolysis remain central questions. Here we use all-atom molecular dynamics simulations of the AAA+ unfoldases Yme1 and Vps4 to explore structural changes induced by hydrolysis, ADP release, and ATP association. The simulations reveal coexisting pre- and post-hydrolysis conformations consistent with experimental cryo- EM densities while also uncovering interconnected allosteric communication networks. One network links the Walker B glutamate to the proximal subunits nucleotide binding pocket (NBP), providing a mechanism for backward propagation of hydrolysis competence around the hexamer. Another network couples NBP and salt bridge rearrangements to pore-loop dynamics, connecting subunit position rather than ATP hydrolysis to substrate release. Simulations of nucleotide exchange intermediates further show that ATP binding before ADP and/or Mg2+ release can induce transient strain leading to alternative hydrolysis-competent configurations that still favor forward progression, providing a structural basis for some of the experimentally observed stochastic behavior. Together, these findings suggest that structural asymmetry creates position-dependent allosteric interactions that generate kinetic asymmetry, biasing a fundamentally stochastic ATPase cycle toward directional substrate translocation while still permitting variation in stepping behavior. Significance StatementAAA+ ATPases play a central role in biology, converting ATP binding, hydrolysis, and release into the directional forces needed to carry out work, such as substrate translocation, remodeling, and degradation. In addition to being important therapeutic targets, understanding how these molecular machines convert stochastic ATP chemistry into persistent mechanical motion is of fundamental interest. Using all-atom molecular dynamics simulations of Yme1 and Vps4, we identify position-responsive conformational changes and allosteric communication networks that propagate hydrolysis competence, coordinate substrate release, and redistribute the relative probabilities of mechanochemical transitions around the ATPase ring. These findings provide molecular insight into how structural asymmetry and kinetic redistribution enable directional substrate translocation.

biophysics↗

bioRxiv2026

Ovarian hormone deficiency enhances wood smoke-induced immune dysfunction via transcriptomic and metabolic alterations

The increasing frequency and severity of wildfires have heightened public exposure to smoke, highlighting the importance of identifying susceptibility factors, including ovarian hormone deficiency. Here, we used single-cell RNA sequencing to profile bone marrow immune cells from ovariectomized (OVX) mice exposed to either filtered air (FA) or wood smoke (WS) followed by functional validation in macrophages from both OVX and Sham-operated mice. Single-cell analyses focused on the OVX context; interactions between surgery and exposure were confirmed at the functional level in assays that included both Sham and OVX groups. In OVX mice, WS broadly suppressed transcriptional programs involved in antigen processing, leukocyte activation, antiviral defense, and bone remodeling. This was associated with altered immune cell composition, including increased memory CD8L T cells and decreased granulocytes and interferon-responsive populations. Bone marrow-derived macrophages (BMDMs) from WS-exposed OVX mice displayed metabolic reprogramming, characterized by the reversal of OVX-induced suppression of oxidative phosphorylation and glycolytic activity, along with reduced expression of M2-associated genes, without concurrent induction of M1-associated genes. This immune-metabolic decoupling suggests that WS exposure under ovarian hormone deficiency may imprint a lasting program in the bone marrow macrophage axis. Together, these findings show that ovarian hormone deficiency increases vulnerability to WS-induced immune disruption in the bone marrow. WS triggers macrophage reprogramming only under ovarian hormone deficiency, leading to heightened metabolic activity alongside suppression of key immune pathways, identifying a novel mechanism of immunotoxicity. These findings emphasize the need to consider hormonal status in air pollution risk assessment.

pharmacology and toxicology↗

bioRxiv2026

Acute wood smoke exposure is associated with cell-specific hippocampal transcriptomic responses in an accelerated ovarian failure mouse model

BackgroundWildfire events are increasing in frequency and intensity, and it is well-known that aging individuals are more susceptible to air pollution exposures, and that air pollution exposures result in neurological sequelae. Despite this, it is unclear how declining levels of ovarian hormones that naturally occur in aging females influence brain vulnerability to air pollution. Menopause and the menopausal transition represent a period of profound physiological change that affects cardiovascular, neurological, and immune health. MethodsWe tested whether perimenopausal-like hormonal status amplifies hippocampal responses to acute wood smoke (WS) using an ovary-intact, 4-vinylcyclohexene diepoxide (VCD) model of moderate accelerated ovarian failure (AOF) in female C57BL/6 mice. Animals were exposed to HEPA-filtered air (FA) or WS for 4 h/day over 2 consecutive days ([~]0.5 mg/m3). Exposure characterization confirmed a complex mixture of combustion products with significant levels of both trace metals and gas release during WS exposure. ResultsSpatial transcriptomics (10x Visium; (n=4 sections/group) with automated cell-type annotation identified astrocytes, GABAergic and glutamatergic neurons, oligodendrocytes, revealed cell type-specific transcriptional alterations following WS exposure. Distinct transcriptional patterns were observed across all identified neuronal and glial cell populations. ConclusionTogether, these findings define a cell type-resolved transcriptional framework linking WS exposure and ovarian hormone decline and identify potential cellular pathways relevant to hippocampal vulnerability.

pharmacology and toxicology↗

bioRxiv2025

LncRNA SChLAP1 promotes cancer cell proliferation and invasion via its distinct structural domains and conserved regions

Long non-coding RNAs (lncRNAs) play key roles in a range of biological processes and disease progression. Despite their functional significance and therapeutic potential, lncRNAs mechanisms of action remain understudied. One such lncRNA is the Second Chromosome Locus Associated with Prostate-1 (SChLAP1). SChLAP1 is overexpressed in malignant prostate cancer and is associated with unfavorable patient outcomes, such as metastasis and increased mortality. In this study, we demonstrated that SChLAP1 possesses distinct structural domains and conserved regions that may contribute to its function. We determined the secondary structure of SChLAP1 using chemical probing methods combined with mutational profiling (DMS-MaP and SHAPE-MaP). Our in vitro secondary structural model revealed that SChLAP1 consists of two distinct secondary-structural modules located at its 5 and 3 ends, both featuring regions with a high degree of structural organization. Our in vivo chemical probing identified potential protein-binding hotspots within the two modules. Overexpression of the modules led to a notable increase in cancer cell proliferation and invasion, proving their functional significance on the oncogenicity of SChLAP1. In conclusion, we discovered functionally important, independent modules with well-defined structures of SChLAP1. These results will serve as a guide to explore the detailed molecular mechanisms by which SChLAP1 promotes aggressive prostate cancer, ultimately contributing to the development of SChLAP1 as a novel therapeutic target.

biochemistry↗

bioRxiv2024

Automated Collective Variable Discovery for MFSD2A transporter from molecular dynamics simulations

Biomolecules often exhibit complex free energy landscapes in which long-lived metastable states are separated by large energy barriers. Overcoming these barriers to robustly sample transitions between the metastable states with classical molecular dynamics (MD) simulations presents a challenge. To circumvent this issue, collective variable (CV)-based enhanced sampling MD approaches are often employed. Traditional CV selection relies on intuition and prior knowledge of the system. This approach introduces bias, which can lead to incomplete mechanistic insights. Thus, automated CV detection is desired to gain a deeper understanding of the system/process. Analysis of MD data with various machine learning algorithms, such as Principal Component Analysis (PCA), Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA)-based approaches have been implemented for automated CV detection. However, their performance has not been systematically evaluated on structurally and mechanistically complex biological systems. Here, we applied these methods to MD simulations of the MFSD2A (Major Facilitator Superfamily Domain 2A) lysolipid transporter in multiple functionally relevant metastable states with the goal of identifying optimal CVs that would structurally discriminate these states. Specific emphasis was on the automated detection and interpretive power of LDA-based CVs. We found that LDA methods, which included a novel gradient descent-based multiclass harmonic variant, termed GDHLDA, we developed here, outperform PCA in class separation, exhibiting remarkable consistency in extracting CVs critical for distinguishing metastable states. Furthermore, the identified CVs included features previously associated with conformational transitions in MFSD2A. Specifically, conformational shifts in transmembrane helix 7 and in residue Y294 on this helix emerged as critical features discriminating the metastable states in MFSD2A. This highlights the effectiveness of LDA-based approaches in automatically extracting from MD trajectories CVs of functional relevance that can be used to drive biased MD simulations to efficiently sample conformational transitions in the molecular system. STATEMENT OF SIGNIFICANCETo elucidate the biological mechanisms of pertinent biomolecules, it is crucial to understand their complex free energy landscapes. Such landscapes are often constructed from molecular dynamics (MD) simulations using collective variable (CV)-guided enhanced sampling methods. Identifying proper CVs for this task is critical but can be challenging with traditional intuition-based approaches. Here we propose an automated protocol for CV discovery which is based on linear discriminant analysis (LDA) for dimensionality reduction of MD data. By applying the protocol to MD simulations of the MFSD2A lysolipid transporter, a structurally and mechanistically complex biological system, we show that LDA-based methods efficiently detect system-specific CVs that accurately classify different metastable states of MFSD2A and are highly interpretable in a detailed structural context.

biophysics↗

bioRxiv2024

Machine Learned Classification of Ligand Intrinsic Activities at Human μ-Opioid Receptor

Opioids are small-molecule agonists of {micro}-opioid receptor ({micro}OR), while reversal agents such as naloxone are antagonists of {micro}OR. Here we developed machine learning (ML) models to classify the intrinsic activities of ligands at the human {micro}OR based on the SMILE strings and two-dimensional molecular descriptors. We first manually curated a database of 983 small molecules with measured Emax values at the human {micro}OR. Analysis of the chemical space allowed identification of dominant scaffolds and structurally similar agonists and antagonists. Decision tree models and directed message passing neural networks (MPNNs) were then trained to classify agonistic and antagonistic ligands. The hold-out test AUCs (areas under the receiver operator curves) of the extra-tree (ET) and MPNN models are 91.5 {+/-} 3.9% and 91.8 {+/-} 4.4%, respectively. To overcome the challenge of small dataset, a student-teacher learning method called tri-training with disagreement was tested using an unlabeled dataset comprised of 15,816 ligands of human, mouse, or rat {micro}OR,{kappa} OR, or{delta} OR. We found that the tri-training scheme was able to increase the hold-out AUC of MPNN to as high as 95.7%. Our work demonstrates the feasibility of developing ML models to accurately predict the intrinsic activities of {micro}OR ligands, even with limited data. We envisage potential applications of these models in evaluating uncharacterized substances for public safety risks and discovering new therapeutic agents to counteract opioid overdoses. TOC Graphic O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=105 SRC="FIGDIR/small/588485v2_ufig1.gif" ALT="Figure 1"> View larger version (16K): org.highwire.dtl.DTLVardef@1bf13f7org.highwire.dtl.DTLVardef@1b7f04aorg.highwire.dtl.DTLVardef@1009569org.highwire.dtl.DTLVardef@1515139_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗

bioRxiv2024

MetaCompare 2.0: Differential ranking of ecological and human health resistome risks

BackgroundWhile there is increasing recognition of numerous environmental contributions to the spread of antibiotic resistance, quantifying the relative contributions of various sources remains a fundamental challenge. Similarly, there is a need to differentiate acute human health risks corresponding to exposure to a given environment, versus broader ecological risk of evolution and spread of antibiotic resistance genes (ARGs) across microbial taxa. Recent studies have proposed various methods of harnessing the rich information housed by metagenomic data for achieving such aims. Here, we introduce MetaCompare 2.0, which improves upon the original MetaCompare pipeline by differentiating indicators of human health resistome risk (i.e., potential for human pathogens to acquire ARGs) from ecological resistome risk (i.e., overall mobility of ARGs across a given microbiome). ResultsTo demonstrate the sensitivity of the MetaCompare 2.0 pipeline, we analyzed publicly available metagenomes representing a broad array of environments, including wastewater, surface water, soil, sediment, and human gut. We also assessed the effect of sequence assembly methods on the risk scores. We further evaluated the robustness of the pipeline to sequencing depth, contig count, and metagenomic library coverage bias through comparative analysis of a range of subsamples extracted from a set of deeply sequenced wastewater metagenomes. The analysis utilizing samples from different environments demonstrated that MetaCompare 2.0 consistently produces lower risk scores for environments with little human influence and higher risk scores for human contaminated environments affected by pollution or other stressors. We found that the ranks of risk scores were not measurably affected by different assemblers employed. The Meta-Compare 2.0 risk scores were remarkably consistent despite varying sequencing depth, contig count, and coverage. ConclusionMetaCompare 2.0 successfully ranked a wide array of environments according to both human health and ecological resistome risks, with both scores being strongly impacted by anthropogenic stress. We packaged the improved pipeline into a publicly-available web service that provides an easy-to-use interface for computing resistome risk scores and visualizing results. The web service is available at http://metacompare.cs.vt.edu/

bioinformatics↗

bioRxiv2024

Detailed mechanisms for unintended large DNA deletions with CRISPR, base editors, and prime editors

CRISPR-Cas9 nucleases are versatile tools for genetic engineering cells and function by producing targeted double-strand breaks (DSBs) in the DNA sequence. However, the unintended production of large deletions (>100 bp) represents a challenge to the effective application of this genome-editing system. We optimized a long-range amplicon sequencing system and developed a k-mer sequence-alignment algorithm to simultaneously detect small DNA alteration events and large DNA deletions. With this workflow, we determined that CRISPR-Cas9 induced large deletions at varying frequencies in cancer cell lines, stem cells, and primary T cells. With CRISPR interference screening, we determined that end resection and the subsequent TMEJ [DNA polymerase theta-mediated end joining] repair process produce most large deletions. Furthermore, base editors and prime editors also generated large deletions despite employing mutated Cas9 "nickases" that produce single-strand breaks. Our findings reveal an important limitation of current genome-editing tools and identify strategies for mitigating unwanted large deletion events.

genetics↗

bioRxiv2023

tICA-Metadynamics for Identifying Slow Dynamics in Membrane Permeation

Molecular simulations are commonly used to understand the mechanism of membrane permeation of small molecules, particularly for biomedical and pharmaceutical applications. However, despite significant advances in computing power and algorithms, calculating an accurate permeation free energy profile remains elusive for many drug molecules because it can require identifying the rate-limiting degrees of freedom (i.e., appropriate reaction coordinates). To resolve this issue, researchers have developed machine learning approaches to identify slow system dynamics. In this work, we apply time-lagged independent component analysis (tICA), an unsupervised dimensionality reduction algorithm, to molecular dynamics simulations with well-tempered metadynamics to find the slowest collective degrees of freedom of the permeation process of trimethoprim through a multicomponent membrane. We show that tICA-metadynamics yields translational and orientational collective variables (CVs) that increase convergence efficiency [~]1.5 times. However, crossing the periodic boundary is shown to introduce artefacts in the translational CV that can be corrected by taking absolute values of molecular features. Additionally, we find that the convergence of the tICA CVs is reached with approximately five membrane crossings, and that data reweighting is required to avoid deviations in the translational CV.

biophysics↗

bioRxiv2023

Darapladib, an inhibitor of Lp-PLA2, sensitizes cancer cells to ferroptosis by remodeling lipid metabolism

Arachidonic and adrenic acids in the membrane play key roles in ferroptosis, but how these fatty acids are manipulated in cells is largely unknown. Here, we reveal that lipoprotein-associated phospholipase A2 (Lp-PLA2) controls intracellular phospholipid metabolism and contributes to ferroptosis resistance. A metabolic drug screen identified that darapladib (SB-480848), an inhibitor of Lp-PLA2, synergistically induced ferroptosis with GPX4 inhibitors. Notably, darapladib was able to enhance ferroptosis under lipoprotein-deficient or serum-free conditions. Furthermore, Lp-PLA2 was located in the membrane and cytoplasm and suppressed ferroptosis, suggesting the critical role of intracellular Lp-PLA2. Lipidomic analysis showed that phosphatidylethanolamine (PE) species were generally enriched, while lysophosphatidylethanolamine (lysoPE) and free fatty acid levels were reduced, upon darapladib treatment. Finally, combination treatment with darapladib and PACMA31, a GPX4 inhibitor, efficiently inhibited tumor growth in a xenograft model. Our study suggests that inhibition of Lp-PLA2 is a potential therapeutic strategy to enhance ferroptosis in cancer treatment.

cancer biology↗

bioRxiv2022

Pancreatic tumors activate arginine biosynthesis to adapt to myeloid-driven amino acid stress

Nutrient stress in the tumor microenvironment requires cancer cells to adopt adaptive metabolic programs to maintain survival and proliferation. Therefore, knowledge of microenvironmental nutrient levels and how cancer cells cope with such nutrition is critical to understand the metabolism underpinning cancer cell biology. Previously, we performed quantitative metabolomics of the interstitial fluid (the local perfusate) of murine pancreatic ductal adenocarcinoma (PDAC) tumors to comprehensively characterize nutrient availability in the microenvironment of these tumors (Sullivan et al., 2019a). Here, we develop Tumor Interstitial Fluid Medium (TIFM), a cell culture medium that contains nutrient levels representative of the PDAC microenvironment, enabling study of PDAC metabolism under physiological nutrition. We show that PDAC cells cultured in TIFM, compared to standard laboratory models, adopt a cellular state more similar to PDAC cells in tumors. Further, using the TIFM model we identified arginine biosynthesis as a metabolic adaptation PDAC cells engage to cope with microenvironmental arginine starvation driven by myeloid cells in PDAC tumors. Altogether, these data show that nutrient availability in tumors is an important determinant of cancer cell metabolism and behavior, and cell culture models that incorporate physiological nutrient availability have improved fidelity and enable the discovery of novel cancer metabolic phenotypes.

cancer biology↗

bioRxiv2021

DeepBioGen: Generalizing predictions to unseen sequencing profiles via visual data augmentation

Predictive models trained on sequencing profiles often fail to achieve expected performance when externally validated on unseen profiles. While many factors such as batch effects, small data sets, and technical errors contribute to the gap between source and unseen data distributions, it is a challenging problem to generalize the predictive models across studies without any prior knowledge of the unseen data distribution. Here, this study proposes DeepBioGen, a sequencing profile augmentation procedure that characterizes visual patterns of sequencing profiles, generates realistic profiles based on a deep generative model capturing the patterns, and generalizes the subsequent classifiers. DeepBioGen outperforms other methods in terms of enhancing the generalizability of the prediction models on unseen data. The generalized classifiers surpass the state-of-the-art method, evaluated on RNA sequencing tumor expression profiles for anti-PD1 therapy response prediction and WGS human gut microbiome profiles for type 2 diabetes diagnosis.

bioinformatics↗

bioRxiv2021

DeepGeni: Deep generalized interpretable autoencoder elucidates gut microbiota for better cancer immunotherapy

Recent studies revealed that gut microbiota modulates the response to cancer immunotherapy and fecal microbiota transplantation has clinical benefit in melanoma patients during the treatment. Understanding microbiota affecting individual response is crucial to advance precision oncology. However, it is challenging to identify the key microbial taxa with limited data as statistical and machine learning models often lose their generalizability. In this study, DeepGeni, a deep generalized interpretable autoencoder, is proposed to improve the generalizability and interpretability of microbiome profiles by augmenting data and by introducing interpretable links in the autoencoder. DeepGeni-based machine learning classifier outperforms state-of-the-art classifier in the microbiome-driven prediction of responsiveness of melanoma patients treated with immune checkpoint inhibitors. DeepGeni-based machine learning classifier outperforms state-of-the-art classifier in the microbiome-driven responsiveness prediction of melanoma patients treated with immune checkpoint inhibitors. Also, the interpretable links of DeepGeni elucidate the most informative microbiota associated with cancer immunotherapy response.

bioinformatics↗

bioRxiv2019

Noncoding de novo mutations contribute to autism spectrum disorder via chromatin interactions

Three-dimensional chromatin structures regulate gene expression across genome. The significance of de novo mutations (DNMs) affecting chromatin interactions in autism spectrum disorder (ASD) remains poorly understood. We generated 931 whole-genome sequences for Korean simplex families to detect DNMs and identified target genes dysregulated by noncoding DNMs via long-range chromatin interactions between regulatory elements. Notably, noncoding DNMs that affect chromatin interactions exhibited transcriptional dysregulation implicated in ASD risks. Correspondingly, target genes were significantly involved in histone modification, prenatal brain development, and pregnancy. Both noncoding and coding DNMs collectively contributed to low IQ in ASD. Indeed, noncoding DNMs resulted in alterations, via chromatin interactions, in target gene expression in primitive neural stem cells derived from human induced pluripotent stem cells from an ASD subject. The emerging neurodevelopmental genes, not previously implicated in ASD, include CTNNA2, GRB10, IKZF1, PDE3B, and BACE1. Our results were reproducible in 517 probands from MSSNG cohort. This work demonstrates that noncoding DNMs contribute to ASD via chromatin interactions.

neuroscience↗

bioRxiv2019

DeepMicro: deep representation learning for disease prediction based on microbiome data

Human microbiota plays a key role in human health and growing evidence supports the potential use of microbiome as a predictor of various diseases. However, the high-dimensionality of microbiome data, often in the order of hundreds of thousands, yet low sample sizes, poses great challenge for machine learning-based prediction algorithms. This imbalance induces the data to be highly sparse, preventing from learning a better prediction model. Also, there has been little work on deep learning applications to microbiome data with a rigorous evaluation scheme. To address these challenges, we propose DeepMicro, a deep representation learning framework allowing for an effective representation of microbiome profiles. DeepMicro successfully transforms high-dimensional microbiome data into a robust low-dimensional representation using various autoencoders and applies machine learning classification algorithms on the learned representation. In disease prediction, DeepMicro outperforms the current best approaches based on the strain-level marker profile in five different datasets. In addition, by significantly reducing the dimensionality of the marker profile, DeepMicro accelerates the model training and hyperparameter optimization procedure with 8X-30X speedup over the basic approach. DeepMicro is freely available at https://github.com/minoh0201/DeepMicro.

bioinformatics↗

bioRxiv2019

A revised 1.6 Å structure of the GTPase domain of the Parkinson’s disease-associated protein LRRK2 provides insights into mechanisms

Leucine-rich repeat kinase 2 (LRRK2) is a large 286 kDa multi-domain protein whose mutation is a common cause of Parkinsons disease (PD). One of the common sites of familial PD-associated mutations occurs at residue Arg-1441 in the GTPase domain of LRRK2. Previously, we reported that the PD-associated mutation R1441H impairs the catalytic activity of the GTPase domain thereby traps it in a persistently \"on\" state. More recently, we reported that the GTPase domain of LRRK2 exists in a dynamic dimer-monomer equilibrium where GTP binding shifts it to the monomeric conformation while GDP binding shifts it back to the dimeric state. We also reported that all of the PD-associated mutations at Arg-1441, including R1441H, R1441C, and R1441G, impair the nucleotide-dependent dimer-monomer conformational dynamics of the GTPase domain. However, the mechanism of this nucleotide-dependent conformational dynamics and how it is impaired by the mutations at residue Arg-1441 remained unclear. Here, we report a 1.6 [A] crystal structure of the GTPase domain of LRRK2. Our structure has revealed a dynamic switch region that can be differentially regulated by GTP and GDP binding. This nucleotide-dependent regulation is impaired when residue Arg-1441 is substituted with the PD-associated mutations due to the loss of its exquisite interactions consisting of two hydrogen bonds and a {pi}-stacking interaction at the dimer interface.\n\nSignificance StatementMutations in LRRK2 are associated with familial Parkinsons disease, so understanding its mechanism of actions and how they are changed by the disease-associated mutations is important for developing therapeutic strategies. This paper describes an atomic structure of the G-domain of LRRK2 revealing that the conformational dynamics of the switch regions are potentially important for its normal function. It further shows that a disease-associated mutation could lock the G domain in a persistently active-like conformation, thus perturbing its normal function.

biochemistry↗