Biology subjects

Kim, J. K.

Publications and source records attributed to Kim, J. K..

At least 55 records · Page 3Linked to original sources

bioRxiv2021

Universally valid reduction of multiscale stochastic biochemical systems using simple non-elementary propensities

Biochemical systems consist of numerous elementary reactions governed by the law of mass action. However, experimentally characterizing all the elementary reactions is nearly impossible. Thus, over a century, their deterministic models that typically contain rapid reversible bindings have been simplified with non-elementary reaction functions (e.g., Michaelis-Menten and Morrison equations). Although the non-elementary reaction functions are derived by applying the quasi-steady-state approximation (QSSA) to deterministic systems, they have also been widely used to derive propensities for stochastic simulations due to computational efficiency and simplicity. However, the validity condition for this heuristic approach has not been identified even for the reversible binding between molecules, such as protein-DNA, enzyme-substrate, and receptor-ligand, which is the basis for living cells. Here, we find that the non-elementary propensities based on the deterministic total QSSA can accurately capture the stochastic dynamics of the reversible binding in general. However, serious errors occur when reactant molecules with similar levels tightly bind, unlike deterministic systems. In that case, the non-elementary propensities distort the stochastic dynamics of a bistable switch in the cell cycle and an oscillator in the circadian clock. Accordingly, we derive alternative non-elementary propensities with the stochastic low-state QSSA, developed in this study. This provides a universally valid framework for simplifying multiscale stochastic biochemical systems with rapid reversible bindings, critical for efficient stochastic simulations of cell signaling and gene regulation. To facilitate the framework, we provide a user-friendly open-source computational package, ASSISTER, that automatically performs the present framework.

systems biology↗

bioRxiv2021

Derivation of stationary distributions of biochemical reaction networks via structure transformation

Long-term behaviors of biochemical reaction networks (BRNs) are described by steady states in deterministic models and stationary distributions in stochastic models. Unlike deterministic steady states, stationary distributions capturing inherent fluctuations of reactions are extremely difficult to derive analytically due to the curse of dimensionality. Here, we develop a method to derive analytic stationary distributions from deterministic steady states by transforming BRNs to have a special dynamic property, called complex balancing. Specifically, we merge nodes and edges of BRNs to match in- and out-flows of each node. This allows us to derive the stationary distributions of a large class of BRNs, including autophosphorylation networks of EGFR, PAK1, and Aurora B kinase and a genetic toggle switch. This reveals the unique properties of their stochastic dynamics such as robustness, sensitivity and multi-modality. Importantly, we provide a user-friendly computational package, CASTANET, that automatically derives symbolic expressions of the stationary distributions of BRNs to understand their long-term stochasticity.

systems biology↗

bioRxiv2021

Inferring causality in biological oscillators

A fundamental goal of biological study is to identify regulatory interactions among components. The recent surge in time-series data collection in biology provides a unique opportunity to infer regulatory networks computationally. However, when the components oscillate, model-free inference methods, while easily implemented, struggle to distinguish periodic synchrony and causality. Alternatively, model-based methods test whether time series are reproducible with a specific model but require inefficient simulations and have limited applicability. Here, we develop an inference method based on a general model of molecular, neuronal, and ecological oscillatory systems that merges the advantages of both model-based and model-free methods, namely accuracy, broad applicability, and usability. Our method successfully infers the positive and negative regulations of various oscillatory networks, including the repressilator and a network of cofactors of pS2 promoter, outperforming popular inference methods. We also provide a computational package, ION (Inferring Oscillatory Networks), that users can easily apply to noisy, oscillatory time series to decipher the mechanisms by which diverse systems generate oscillations.

systems biology↗

bioRxiv2021

Personalized sleep-wake patterns aligned with circadian rhythm relieve daytime sleepiness

Shift workers and many other groups experience irregular sleep-wake patterns. This can induce excessive daytime sleepiness that decreases productivity and elevates the risk of accidents. However, the degree of daytime sleepiness is not correlated with standard sleep parameters like total sleep time, suggesting other factors are involved. Here, we analyze real-world sleep-wake patterns of shift workers measured by wearables with a newly developed user-friendly computational package that simulates homeostatic sleep pressure - the physiological need for sleep - and the circadian rhythm. This reveals that shift workers who align sleep-wake patterns with their circadian rhythm have lower daytime sleepiness, even if they sleep less. The alignment, quantified by a new parameter, circadian sleep sufficiency, can be increased by dynamically adjusting daily sleep durations according to varying bedtimes. Our computational package provides flexible and personalized real-time sleep-wake patterns for individuals to reduce their daytime sleepiness and could be used with wearable devices to develop smart alarms.

systems biology↗

bioRxiv2020

CRISPR-enhanced human adipocyte 'browning' as cell therapy for metabolic disease

Obesity and type 2 diabetes (T2D) are associated with poor tissue responses to insulin1,2, disturbances in glucose and lipid fluxes3-5 and comorbidities including steatohepatitis6 and cardiovascular disease7,8. Despite extensive efforts at prevention and treatment9,10, diabetes afflicts over 400 million people worldwide11. Whole body metabolism is regulated by adipose tissue depots12-14, which include both lipid-storing white adipocytes and less abundant "brown" and "brite/beige" adipocytes that express thermogenic uncoupling protein UCP1 and secrete factors favorable to metabolic health15-18. Application of clustered regularly interspaced short palindromic repeats (CRISPR) gene editing19,20 to enhance "browning" of white adipose tissue is an attractive therapeutic approach to T2D. However, the problems of cell-selective delivery, immunogenicity of CRISPR reagents and long term stability of the modified adipocytes are formidable. To overcome these issues, we developed methods that deliver complexes of SpyCas9 protein and sgRNA ex vivo to disrupt the thermogenesis suppressor gene NRIP121,22 with near 100% efficiency in human or mouse adipocytes. NRIP1 gene disruption at discrete loci strongly ablated NRIP1 protein and upregulated expression of UCP1 and beneficial secreted factors, while residual Cas9 protein and sgRNA were rapidly degraded. Implantation of the CRISPR-enhanced human or mouse brown-like adipocytes into high fat diet fed mice decreased adiposity and liver triglycerides while enhancing glucose tolerance compared to mice implanted with unmodified adipocytes. These findings advance a therapeutic strategy to improve metabolic homeostasis through CRISPR-based genetic modification of human adipocytes without exposure of the recipient to immunogenic Cas9 or delivery vectors.

bioengineering↗

bioRxiv2020

Characterization of Viral Insulins Reveals White Adipose Tissue Specific Effects in Mice

Members of the insulin/IGF superfamily are well conserved across the evolutionary tree. We recently showed that four viruses in the Iridoviridae family possess genes that encode proteins highly homologous to human insulin/IGF-1. Using chemically synthesized single chain (sc), i.e. IGF-1-like, forms of the viral insulin/IGF-1 like peptides (VILPs), we previously showed that they can stimulate human receptors. Because these peptides possess potential cleavage sites to form double chain (dc), i.e. more insulin-like, VILPs, in this study, we have characterized dc forms of VILPs for Grouper iridovirus (GIV), Singapore grouper iridovirus (SGIV) and Lymphocystis disease virus-1 (LCDV-1). GIV and SGIV dcVILPs bind to both isoforms of human insulin receptor (IR-A, IR-B) and to the IGF1R, and for the latter show higher affinity than human insulin. These dcVILPs stimulate IR and IGF1R phosphorylation and post-receptor signaling in vitro and in vivo. Both GIV and SGIV dcVILPs stimulate glucose uptake in mice. In vivo infusion experiments in awake mice revealed that while insulin (0.015 nmol/kg/min) and GIV dcVILP (0.75nmol/kg/min) stimulated a comparable glucose uptake in heart, skeletal muscle and brown adipose tissue, GIV dcVILP stimulated ~2 fold higher glucose uptake in white adipose tissue (WAT) compared to insulin. This was associated with increased Akt phosphorylation and glucose transporter type 4 (GLUT4) gene expression compared to insulin. Taken together, these results show that GIV and SGIV dcVILPs are active members of the insulin superfamily with unique characteristics. Elucidating the mechanism of tissue specificity for GIV dcVILP will help us to better understand insulin action, design new analogues that specifically target the tissues, and provide new insights into their potential role in disease.

cell biology↗

bioRxiv2020

SNPC-1.3 is a sex-specific transcription factor that drives male piRNA expression in C. elegans

Piwi-interacting RNAs (piRNAs) play essential roles in silencing repetitive elements to promote fertility in metazoans. Studies in worms, flies, and mammals reveal that piRNAs are expressed in a sex-specific manner. However, the mechanisms underlying this sex-specific regulation are unknown. Here we identify SNPC-1.3, a variant of a conserved subunit of the snRNA activating protein complex, as a male-specific piRNA transcription factor in C. elegans. Binding of SNPC-1.3 at male piRNA loci drives spermatogenic piRNA transcription and requires the core piRNA transcription factor SNPC-4. Loss of snpc-1.3 leads to depletion of male piRNAs and defects in male-dependent fertility. Furthermore, TRA-1, a master regulator of sex determination, binds to the snpc-1.3 promoter and represses its expression during oogenesis. Loss of TRA-1 targeting causes ectopic expression of snpc-1.3 and male piRNAs during oogenesis. Thus, sexual dimorphic regulation of snpc-1.3 coordinates male and female piRNA expression during germline development.

developmental biology↗

bioRxiv2019

Mutation of a PER2 phosphodegron perturbs the circadian phosphoswitch

Casein kinase 1 (CK1) plays a central role in regulating the period of the circadian clock. In mammals, PER2 protein abundance is regulated by CK1-mediated phosphorylation and proteasomal degradation. On the other hand, recent studies have questioned whether the degradation of the core circadian machinery is a critical step in clock regulation. Prior cell-based studies found that CK1 phosphorylation of PER2 at Ser478 recruits the ubiquitin E3 ligase {beta}-TrCP, leading to PER2 degradation. Creation of this phosphodegron is regulated by a phosphoswitch that is also implicated in temperature compensation. However, in vivo evidence that this phosphodegron influences circadian period is lacking. Here, we generated and analyzed PER2-Ser478Ala knock-in mice. The mice showed longer circadian period in behavioral analysis. Molecularly, mutant PER2 protein accumulated in both the nucleus and cytoplasm of the mouse liver, while Per2 mRNA levels were minimally affected. Nuclear PER1, CRY1 and CRY2 proteins also increased, probably due to stabilization of PER2-containing complexes. In mouse embryonic fibroblasts derived from PER2-Ser478Ala::LUC mice, three-phase decay and temperature compensation of the circadian period was perturbed. These data provide direct in vivo evidence for the importance of phosphorylation-regulated PER2 stability in the circadian clock and validate the phosphoswitch in a mouse model.

molecular biology↗

bioRxiv2019

A big-data approach to understanding metabolic rate and response to obesity in laboratory mice

Maintaining a healthy body weight requires an exquisite balance between energy intake and energy expenditure. In humans and in laboratory mice these factors are experimentally measured by powerful and sensitive indirect calorimetry devices. To understand the genetic and environmental factors that contribute to the regulation of body weight, an important first step is to establish the normal range of metabolic values and primary sources contributing to variability in results. Here we examine indirect calorimetry results from two experimental mouse projects, the Mouse Metabolic Phenotyping Centers and International Mouse Phenotyping Consortium to develop insights into large-scale trends in mammalian metabolism. Analysis of nearly 10,000 wildtype mice revealed that the largest experimental variances are consequences of institutional site. This institutional effect on variation eclipsed those of housing temperature, body mass, locomotor activity, sex, or season. We do not find support for the claim that female mice have greater metabolic variation than male mice. An analysis of these factors shows a normal distribution for energy expenditure in the phenotypic analysis of 2,246 knockout strains and establishes a reference for the magnitude of metabolic changes. Using this framework, we examine knockout strains with known metabolic phenotypes. We compare these effects with common environmental challenges including age, and exercise. We further examine the distribution of metabolic phenotypes exhibited by knockout strains of genes corresponding to GWAS obesity susceptibility loci. Based on these findings, we provide suggestions for how best to design and conduct energy balance experiments in rodents, as well as how to analyze and report data from these studies. These recommendations will move us closer to the goal of a centralized physiological repository to foster transparency, rigor and reproducibility in metabolic physiology experimentation.

physiology↗

bioRxiv2019

Ribosome dimerization prevents loss of essential ribosomal proteins during quiescence

The formation of ribosome dimers during periods of quiescence is widespread among bacteria and some higher eukaryotes. However, the mechanistic importance of dimerization is not well understood. In bacteria ribosome dimerization is mediated by the Hibernation Promoting Factor (HPF). Here, we report that HPF from the Gram-positive bacterium Bacillus subtilis preserves active ribosomes by preventing the loss of essential ribosomal proteins. Ribosomes isolated from strains either lacking HPF ({Delta}hpf) or encoding a mutant allele of HPF that binds the ribosome but does not mediate dimerization were substantially depleted of the small subunit proteins S2 and S3. Strikingly, these proteins are located at the ribosome dimer interface. We used single particle cryo-EM to further characterize ribosomes isolated from a {Delta}hpf mutant strain and observed that many were missing S2, S3, or both. These data support a model in which the ribosome dimerization activity of HPF evolved to protect labile proteins that are essential for ribosome function.\n\nSignificance StatementWhen nutrients become scarce, many bacterial species enter an extended state of quiescence. A major challenge of this state is how to attenuate protein synthesis, the most energy consuming cellular process, while preserving ribosomes for the return to favorable conditions. Here, we show that the ribosome-binding protein HPF which dimerizes ribosomes functions to protect essential ribosomal proteins at the dimer interface. HPF is almost universally conserved in bacteria and HPF deletions in diverse species exhibit decreased viability under nutrient limitation. Our data provide mechanistic insight into this phenotype and establish a role for HPF in maintaining translationally competent ribosomes during quiescence.

microbiology↗

bioRxiv2019

Soft Windowing Application to Improve Analysis of High-throughput Phenotyping Data

MotivationHigh-throughput phenomic projects generate complex data from small treatment and large control groups that increase the power of the analyses but introduce variation over time. A method is needed to utlize a set of temporally local controls that maximises analytic power while minimising noise from unspecified environmental factors.\n\nResultsHere we introduce \"soft windowing\", a methodological approach that selects a window of time that includes the most appropriate controls for analysis. Using phenotype data from the International Mouse Phenotyping Consortium (IMPC), adaptive windows were applied such that control data collected proximally to mutants were assigned the maximal weight, while data collected earlier or later had less weight. We applied this method to IMPC data and compared the results with those obtained from a standard non-windowed approach. Validation was performed using a resampling approach in which we demonstrate a 10% reduction of false positives from 2.5 million analyses. We applied the method to our production analysis pipeline that establishes genotype-phenotype associations by comparing mutant versus control data. We report an increase of 30% in significant p-values, as well as linkage to 106 versus 99 disease models via phenotype overlap with the soft windowed and non-windowed approaches, respectively, from a set of 2,082 mutant mouse lines. Our method is generalisable and can benefit large-scale human phenomic projects such as the UK Biobank and the All of Us resources.\n\nAvailability and ImplementationThe method is freely available in the R package SmoothWin, available on CRAN http://CRAN.R-project.org/package=SmoothWin.

bioinformatics↗

bioRxiv2019

RNA promotes phase separation of glycolysis enzymes into yeast G bodies in hypoxia

In hypoxic stress conditions, glycolysis enzymes assemble into singular cytoplasmic granules called glycolytic (G) bodies. Formation of G bodies in yeast is correlated with increased glucose consumption and cell survival. However, the physical properties and organizing principles that define G body formation are unclear. We demonstrate that glycolysis enzymes are non-canonical RNA binding proteins, sharing many common mRNA substrates that are also integral constituents of G bodies. Tethering a G body component, the beta subunit of the yeast phosphofructokinase, Pfk2, to nonspecific endoribonucleases reveals that RNA nucleates G body formation and subsequent maintenance of G body structural integrity. Consistent with a phase separation mechanism of G body formation, recruitment of glycolysis enzymes to G bodies relies on multivalent homotypic and heterotypic interactions. Furthermore, G bodies can fuse in live cells and are largely insensitive to 1,6-hexanediol treatment, consistent with a hydrogel-like state in its composition. Taken together, our results elucidate the biophysical nature of G bodies and demonstrate that RNA nucleates phase separation of the glycolysis machinery in response to hypoxic stress.

cell biology↗

bioRxiv2019

Bayesian inference of distributed time delay in transcriptional and translational regulation

MotivationAdvances in experimental and imaging techniques have allowed for unprecedented insights into the dynamical processes within individual cells. However, many facets of intracellular dynamics remain hidden, or can be measured only indirectly. This makes it challenging to reconstruct the regulatory networks that govern the biochemical processes underlying various cell functions. Current estimation techniques for inferring reaction rates frequently rely on marginalization over unobserved processes and states. Even in simple systems this approach can be computationally challenging, and can lead to large uncertainties and lack of robustness in parameter estimates. Therefore we will require alternative approaches to efficiently uncover the interactions in complex biochemical networks.\n\nResultsWe propose a Bayesian inference framework based on replacing uninteresting or unobserved reactions with time delays. Although the resulting models are non-Markovian, recent results on stochastic systems with random delays allow us to rigorously obtain expressions for the likelihoods of model parameters. In turn, this allows us to extend MCMC methods to efficiently estimate reaction rates, and delay distribution parameters, from single-cell assays. We illustrate the advantages, and potential pitfalls, of the approach using a birth-death model with both synthetic and experimental data, and show that we can robustly infer model parameters using a relatively small number of measurements. We demonstrate how to do so even when only the relative molecule count within the cell is measured, as in the case of fluorescence microscopy.

systems biology↗

bioRxiv2019

The full-length transcriptome of C. elegans using direct RNA sequencing

Current transcriptome annotations have largely relied on short read lengths intrinsic to most widely used high-throughput cDNA sequencing technologies. For example, in the annotation of the Caenorhabditis elegans transcriptome, more than half of the transcript isoforms lack full-length support and instead rely on inference from short reads that do not span the full length of the isoform. We applied nanopore-based direct RNA sequencing to characterize the developmental polyadenylated transcriptome of C. elegans. Taking advantage of long reads spanning the full length of mRNA transcripts, we provide support for 20,902 splice isoforms across 14,115 genes, without the need for computational reconstruction of gene models. Of the isoforms identified, 2,188 are novel splice isoforms not present in the Wormbase WS265 annotation. Furthermore, we identified 16,325 3 untranslated region (3UTR) isoforms, 2,304 of which are novel and do not fall within 10 bp of existing 3UTR datasets and annotations. Combining 3UTRs and splice isoforms we identified 25,944 full-length isoforms. We also determined that poly(A) tail lengths of transcripts vary across development, as do the strengths of previously reported correlations between poly(A) tail length and expression level, and poly(A) tail length and 3UTR length. Finally, we have formatted this data as a publically accessible track hub, enabling researchers to explore this dataset easily in a genome browser.

genomics↗

bioRxiv2019

Physicochemical Properties of Chromosomes in Live Cells Characterized by Label-Free Imaging and Fluorescence Correlation Spectroscopy

The cell nucleus is a three-dimensional, dynamic organelle that is organized into many subnuclear bodies, such as chromatin and nucleoli. The structure and function of these bodies is maintained by diffusion and interactions between related factors as well as dynamic and structural changes. Recent studies using fluorescent microscopic techniques suggest that protein factors can access and are freely mobile in mitotic chromosomes, despite their densely packed structure. However, the physicochemical properties of the chromosome itself during cell division are not yet fully understood. Physical parameters, such as the refractive index (RI), volume of the mitotic chromosome, and diffusion coefficients of fluorescent probes inside the chromosome were quantified using an approach combining label-free optical diffraction tomography with complementary confocal laser scanning microscopy and fluorescence correlation spectroscopy. Variance in these parameters correlated among various osmotic conditions, suggesting that changes in RI are consistent with those in the diffusion coefficient for mitotic chromosomes and cytosol. Serial RI tomography images of chromosomes in live cells during mitosis were compared with three-dimensional confocal micrographs to demonstrate that compaction and decompaction of chromosomes induced by osmotic change were characterized by linked changes in chromosome RI, volume, and the mobility of fluorescent proteins.

biophysics↗

bioRxiv2018

Resistance to Diet Induced Obesity in the Apolipoprotein E Deficient Mouse is Associated with an Attenuated Transcriptional Response in Visceral Fat

The apolipoprotein E knockout (EKO) mouse is a well-established model of atherosclerosis. Macrophages in the arterial intima of EKO mice serve a protective role, scavenging oxidatively modified LDL in order to protect cells from toxic free cholesterol. Recent studies have highlighted a similar role for macrophage foam cells in restraining the increased rates of lipolysis in adipose tissue of obese and fasting mice. Interestingly, EKO adipocytes have been shown to have increased rates of lipolysis in vitro. Therefore, the aim of this study was to examine how apoE deficiency might alter the transcriptional response of visceral adipose tissue (VAT) to high fat diet (HFD). EKO mice fed HFD for 24 weeks gained less fat mass and were more insulin sensitive than their wild type (WT) littermates. Metabolic cages showed that HFD EKO mice had increased post-prandial oxygen consumption and increased serum {beta}-hydroxybutyrate. DNA microarrays revealed that EKO VAT was comparatively insensitive to HFD in terms of alterations in gene expression, with only 0.1% of probe sets differentially expressed. In contrast, the VAT of WT mice had a 30 fold more extensive alteration in gene expression (3% of probes sets), characterized predominantly by increased expression of immune cell specific genes. In addition, analysis of a priori determined gene sets revealed broad down-regulation of PPAR{gamma} target and fatty acid catabolism genes in WT VAT, and increased expression of lipid storage and cholesterol synthesis genes. In comparison, expression of PPAR{gamma} target genes was not down-regulated in EKO VAT and expression of fatty acid oxidation genes was increased. In summary, we report three novel findings with regards to metabolism in the EKO mouse: 1) increased post-prandial oxygen consumption, 2) increased serum {beta} hydroxybutyrate concentrations and 3) a dramatically less robust transcriptional response to HFD in EKO VAT. These findings suggest that limiting adipocyte exposure to dietary fatty acids may be an attractive therapy for diet induced obesity, provided that compensatory mechanisms that prevent hyperlipidemia can be activated.

physiology↗

bioRxiv2018

A kinesin Klp10A mediates cell cycle-dependent shuttling of Piwi between nucleus and nuage

The piRNA pathway protects germline genomes through transcript cleavage of selfish genetic elements, such as transposons, in the cytoplasm and their transcriptional silencing in the nucleus. Here, we describe a mechanism by which the nuclear and cytoplasmic arms of the silencing mechanism are linked. During mitosis of Drosophila spermatogonia, nuclear Piwi interacts with nuage, the compartment that mediates the cytoplasmic arm of piRNA-mediated silencing. At the end of mitosis, Piwi leaves nuage to return to the nucleus. We found that dissociation of Piwi from nuage occurs at the depolymerizing microtubules of the central spindle, mediated by a microtubule-depolymerizing kinesin Klp10A. Depletion of klp10A delays Piwis return to the nucleus and affects piRNA production, suggesting the importance of nuclear-cytoplasmic communication in piRNA biogenesis. We propose that cell cycle-dependent communication between the nuclear and cytoplasmic arms of the piRNA pathway plays important roles in coordinated piRNA production.

developmental biology↗

bioRxiv2018

CRISPR delivery particles for developing therapeutic strategies in metabolic disease

RNA-guided engineered nucleases derived from a prokaryotic adaptive immune system known as CRISPR-Cas represent a promising platform for gene deletion and editing. As a therapeutic approach, direct delivery of Cas9 protein and guide RNA could circumvent the safety problems associated with plasmid delivery and therefore represents an attractive tool for genome engineering. Gene deletion or editing in adipose tissue to enhance its energy expenditure, fat oxidation and secretion of bioactive factors through a \"browning\" process presents a potential therapeutic strategy to alleviate metabolic disease. Here, we developed novel CRISPR delivery particles, denoted CriPs, composed of nano-size complexes of Cas9 protein and single guide (sg)RNA, coated with an amphipathic peptide called Endo-Porter that mediates entry into cells. Efficient CRISPR-Cas9 mediated gene deletion of ectopically expressed Green fluorescent protein (GFP) by CriPs was achieved in multiple cell types including a macrophage cell line, primary macrophages and primary pre-adipocytes. Significant GFP loss was also observed in peritoneal exudate cells with minimum systemic toxicity in GFP expressing mice following intraperitoneal injection of CriPs containing sgRNA targeting Gfp. Furthermore, the disruption of the Nrip1 gene in white adipocytes by CriPs enhanced adipocyte \"browning\" with a marked increase of UCP1 expression. Deletion of Nrip1 by CriPs did not produce detectable off-target effects. Thus CriPs represent a novel CRISPR delivery system for Cas9 and sgRNA that is effective for ablating targeted gene products in cultured cells and in vivo, and provide a potential therapeutic strategy for metabolic disease.

bioengineering↗