Biology subjects

Lu, H.

Publications and source records attributed to Lu, H..

At least 145 records · Page 8Linked to original sources

bioRxiv2023

Effects of Probiotics, Glucose Oxidase and its Combination on Immune Function, Antioxidant Status, Serum Biochemical Incides and Toxin Residue in Sheep when Challenged with Aflatoxin B1

The objective of this study was to investigate the ability of a mixed mycotoxins detoxification agent (probiotics, Glucose oxidase (GOD) and its combination) to alleviate the toxicity of aflatoxin B1 (AFB1) by assessing growth performance, serum toxin residue, immune function, antioxidant status and serum biochemical indices of sheep. Twenty 8-month-old Mongolian sheep were randomly assigned to 1-5 diet treatment groups: 1) the control (CON) group: basal diet; 2) the AFB1 (A) group: CON + 0.2 mg/kg AFB1; 3) the AFB1 and probiotics (AP) group: A + 0.5% probiotics; 4) the AFB1 and GOD (AG) group: A + 0.5% GOD; 5) the AFB1, probiotics, and GOD (APG) group: A + 0.5% probiotics + 0.5% GOD. Results showed that treatment A adversely affected the health and performance of sheep. However, the treatments AP, AG, or APG exerted a positive effect on health, performance and other indications. In conclusion, probiotics, GOD, and its combination induced injury of Mongolian sheep by alleviating the negative effects of AFB1 on the production performance, immune indexes, antioxidant indexes, and serum biochemical indexes and effectively reducing serum toxin residue. Key ContributionThe study investigated the effects of probiotics, Glucose oxidase and its combination on serum detoxification of AFB1 by assessing growth performance, serum toxin residue, immune function, antioxidant status and blood biochemical indices of sheep. AFB1 - contaminated diets supplemented with probiotics and GOD were effective in improving growth performance, immunity, antioxidant function, liver function, and decreasing serum AFB1 residue of sheep.

zoology↗

bioRxiv2023

Dissecting the Impact of Maternal Androgen Exposure on Offspring Health through Targeting the Androgen Receptor in Developmental Programming

Women with polycystic ovary syndrome (PCOS) exhibit sustained elevation in circulating androgens during pregnancy, an independent risk factor linked to pregnancy complications and adverse neonatal outcomes. Yet, further investigation is required to understand the precise mechanisms and the impact on cell-type specific placental dysfunction. To explore these dynamics, a PCOS-like mice model was induced with continuous androgen exposure throughout pregnancy, mimicking the human-PCOS. This resulted in impaired placental and embryonic development, leading to mid-gestation lethality. Co-treatment with the androgen receptor blocker, flutamide, prevented this lethality. Comprehensive analysis using whole-genome bisulfite and RNA sequencing revealed the diminished proportion of trophoblast precursors by downregulation of Cdx2. The absence of Gcm1, Synb, and Prl3b1 further resulted in decreased numbers of syncytiotrophoblasts and sinusoidal trophoblast giant cells, leading to observed compromised placenta labyrinth formation. Importantly, human trophoblast organoids exposed to androgens exhibited analogous alterations, highlighting impaired trophoblast differentiation as a key feature in PCOS-related pregnancy complications. Remarkably, all effects were mediated through the androgen receptor pathways, as demonstrated by comparable offspring phenotypes to controls when treated with flutamide. These findings provide novel insight into the PCOS-related pregnancy complications, and potential cellular targets for future treatment.

physiology↗

bioRxiv2023

DeepEnzyme: a robust deep learning model for improved enzyme turnover number prediction by utilizing features of protein 3D Structures

Turnover numbers (kcat), which indicate an enzymes catalytic efficiency, have a wide range of applications in fields including protein engineering and synthetic biology. Experimentally measuring the enzymes kcat is always time-consuming. Recently, the prediction of kcat using deep learning models has mitigated this problem. However, the accuracy and robustness in kcat prediction still needs to be improved significantly, particularly when dealing with enzymes with low sequence similarity compared to those within the training dataset. Herein, we present DeepEnzyme, a cutting-edge deep learning model that combines the most recent Transformer and Graph Convolutional Network (GCN) architectures. To improve the prediction accuracy, DeepEnzyme was trained by leveraging the integrated features from both sequences and 3D structures. Consequently, our model exhibits remarkable robustness when processing enzymes with low sequence similarity compared to those in the training dataset by utilizing additional features from high-quality protein 3D structures. DeepEnzyme also makes it possible to evaluate how point mutations affect the catalytic activity of the enzyme, which helps identify residue sites that are crucial for the catalytic function. In summary, DeepEnzyme represents a pioneering effort in predicting enzymes kcat values with superior accuracy and robustness compared to previous algorithms. This advancement will significantly contribute to our comprehension of enzyme function and its evolutionary patterns across species.

bioinformatics↗

bioRxiv2023

Yeast9: A Consensus Yeast Metabolic Model Enables Quantitative Analysis of Cellular Metabolism By Incorporating Big Data

Genome-scale metabolic models (GEMs) can facilitate metabolism-focused multi-omics integrative analysis. Since Yeast8, the yeast-GEM of Saccharomyces cerevisiae, published in 2019, has been continuously updated by the community. This have increased the quality and scope of this model, culminating now in Yeast9. To evaluate its predictive performance, we generated 163 condition-specific GEMs constrained by single-cell transcriptomics from osmotic pressure or normal conditions. Comparative flux analysis showed that yeast adapting to high osmotic pressure benefits from upregulating fluxes through the central carbon metabolism. Furthermore, combining Yeast9 with proteomics revealed metabolic rewiring underlying its preference in nitrogen sources. Lastly, we created strain-specific GEMs (ssGEMs) constrained by transcriptomics for 1229 mutant strains. Well able to predict the strains growth rates, fluxomics from those large-scale ssGEMs outperformed transcriptomics in predicting functional categories for all studied genes in machine-learning models. Based on those findings we anticipate that Yeast9 will empower systems biology studies of yeast metabolism.

systems biology↗

bioRxiv2023

High-throughput experimental evolution reveals the mutation-driven landscape of high-level clinical resistance acquired under low-level environmental antibiotic stresses

Potential pathogens consistently exposed to low-level environmental antibiotics could derive high-level clinically relevant resistance detrimental to the human health. However, the underlying evolutionary landscapes remain poorly understood. We conducted a high-throughput experimental evolution study by exposing an environmentally isolated pathogenic Escherichia coli strain to 96 typical antibiotics at 10 g l-1 for 200 generations. Antibiotic resistance phenotypic (IC90 against 8 clinically used antibiotics) and genetic changes of the evolved populations were systematically investigated, revealing a universal increase in antibiotic resistance (up to 349-fold), and mutations in 2,432 genes. Transposon sequencing was further employed to verify genes potentially associated with resistance. A core set of mutant genes conferring high-level resistance was analyzed to elucidate their resistance mechanisms by analyzing the functions of interacted genes within the gene co-fitness network and performing gene knockout validations. We developed machine-learning models to predict antibiotic resistance phenotypes from antibiotic structures and genomic mutations, enabling the resistance predictions for another 569 antibiotics. Importantly, 14.6% of the 481 key mutations were observed in clinical and environmental E. coli isolates retrieved from the NCBI database, and several were over-represented in >500 clinical isolates. Deciphering the evolutionary landscapes underlying resistance exposed to low-level environmental antibiotics is crucial for evaluating the emergence and risks of environment-originated clinical antibiotic resistance.

evolutionary biology↗

bioRxiv2023

Transfer of disulfide bond formation modules via yeast artificial chromosomes promotes the expression of heterologous proteins in Kluyveromyces marxianus

Kluyveromyces marxianus is a food-safe yeast with great potential for producing heterologous proteins. Improving the yield in K. marxianus remains a challenge, while incorporating large-scale functional modules poses a technical obstacle in engineering. To address these issues, linear and circular yeast artificial chromosomes of K. marxianus (KmYACs) were constructed and loaded with disulfide bond formation modules from Pichia pastoris or K. marxianus. These modules contained up to 7 genes with a maximum size of 15 kb. KmYACs carried telomeres either from K. marxianus or Tetrahymena. KmYACs were transferred successfully into K. marxianus and stably propagated without affecting the normal growth of the host, regardless of the type of telomeres and configurations of KmYACs. KmYACs increased the overall expressions of disulfide bond formation genes and significantly enhanced the yield of various heterologous proteins. In high-density fermentation, the use of KmYACs resulted in a glucoamylase yield of 16.8 g/L, the highest reported level to date in K. marxianus. Transcriptomic and metabolomic analysis of cells containing KmYACs suggested increased FAD biosynthesis, enhanced flux entering the TCA cycle and a preferred demand for lysine and arginine as features of cells overexpressing heterologous proteins. Consistently, supplementing lysine or arginine further improved the yield. Therefore, KmYAC provides a powerful platform for manipulating large modules with enormous potential for industrial applications and fundamental research. Transferring the disulfide bond formation module via YACs proves to be an efficient strategy for improving the yield of heterologous proteins, and this strategy may be applied to optimize other microbial cell factories. Impact StatementIn this study, yeast artificial chromosomes of K. marxianus (KmYACs) were constructed and successfully incorporating modules for large-scale disulfide bond formation. KmYACs were stably propagated in K. marxianus without compromising the normal growth of the host, irrespective of the selection of telomeres (either Tetrahymena or K. marxianus) and configuration (either linear or circular). KmYACs notably enhanced the expressions of various heterologous proteins, with further yield improvement by supplementing lysine or arginine in the medium. Our findings affirm KmYAC as a robust and versatile platform for transferring large-scale function modules, demonstrating immense potential for both industrial applications and fundamental research.

synthetic biology↗

bioRxiv2023

Chaperone-mediated autophagy is an overlooked pathway for mutant α1 antitrypsin Z degradation

Chaperone-mediated autophagy (CMA) is a specific form of autophagy that selectively targets proteins containing a KFERQ-like motif and relies on the chaperone protein HSC70 for substrate recognition. In 1-antitrypsin deficiency (AATD), a disease characterized by the hepatic build-up of 1-Antitrypsin Z mutant (ATZ), CMAs role had been unclear. This work demonstrates the critical role that CMA plays in preventing ATZ accumulation; suppressing CMA worsens ATZ accumulation, whilst activating it through chemical stimulation or LAMP2A overexpression promotes ATZ breakdown. Specifically, ATZs 121QELLR125 motif is critical for HSC70 recognition and LAMP2As charged C-terminal cytoplasmic tail is vital for substrate binding, facilitating CMA-mediated degradation of ATZ. This selective activation of CMA operates independently from other autophagy pathways and alleviate ATZ aggregates caused cellular stress. These findings highlight CMAs critical function in cellular protein quality control of ATZ and place it as a novel target for AATD treatment approaches.

cell biology↗

bioRxiv2023

Bacteriophages from human skin infecting coagulase-negative Staphylococcus: diversity, novel species and host resistance

The human skin microbiome comprises diverse populations that differ temporally between body sites and individuals. The virome is a less studied component of the skin microbiome and the study of bacteriophages is required to increase knowledge of the modulation and stability of bacterial communities. Staphylococcus species are among the most abundant colonisers of skin and are associated with both health and disease yet the bacteriophages infecting the most abundant species on skin are less well studied. Here, we report the isolation and genome sequencing of 40 bacteriophages from human skin swabs that infect coagulase-negative Staphylococcus (CoNS) species, which extends our knowledge of phage diversity. Six genetic clusters of phages were identified with two clusters representing novel phage species, one of which we characterise and name Alsa phage. We identified that Alsa phages have a greater ability to infect the species S. hominis that was otherwise infected less than other CoNS species by the isolated phages, indicating an undescribed barrier to phage infection that could be in part due to numerous restriction-modification systems. The extended diversity of Staphylococcus phages here enables further research to define their contribution to skin microbiome research and the mechanisms that limit phage infection.

microbiology↗

bioRxiv2023

Liver-specific glucocorticoid action in alcoholic liver disease: study of glucocorticoid receptor knockout and knockin mice

BackgroundGlucocorticoids are the only first-line drugs for severe alcoholic hepatitis (AH), with limited efficacy and various side effects on extrahepatic tissues. Liver-targeting glucocorticoid therapy may have multiple advantages over systemic glucocorticoid for AH. The purpose of this study was to determine the role of hepatocellular glucocorticoid receptor (GR) in alcoholic steatosis (AS) and AH. MethodsAS was induced by a high-fat diet plus binge alcohol in adult male and female mice with liver-specific knockout (LKO) and heterozygote of GR. AH was induced by chronic-plus-binge in middle-aged male mice with liver-specific knockin of GR. Changes in hepatic mRNA and protein expression were determined by qPCR and Western blot. ResultsGR LKO aggravated steatosis and decreased hepatic expression and circulating levels of albumin in both genders of AS mice but only increased markers of liver injury in male AS mice. Marked steatosis in GR LKO mice was associated with induction of lipogenic genes and down-regulation of bile acid synthetic genes. Hepatic protein levels of GR, hepatocyte nuclear factor 4, and phosphorylated STAT3 were gene-dosage-dependently decreased, whereas that of lipogenic ATP citrate lyase was increased in male GR heterozygote and LKO mice. Interestingly, hepatic expression of estrogen receptor (ER) was induced, and the essential estrogen-inactivating enzyme sulfotransferase 1E1 was gene-dosage-dependently down-regulated in GR heterozygote and knockout AS mice, which was associated with induction of ER-target genes. Liver-specific knockin of GR protected against liver injury and steatohepatitis in middle-aged AH mice. ConclusionsHepatocellular GR is important for protection against AS and AH.

pharmacology and toxicology↗

bioRxiv2023

HILAMA: High-dimensional multi-omic mediation analysis with latent confounding

MotivationThe increasingly available multi-omic datasets have posed both new opportunities and challenges to the development of quantitative methods for discovering novel mechanisms in biomedical research. One natural approach to analyzing such datasets is mediation analysis originated from the causal inference literature. Mediation analysis can help unravel the mechanisms through which exposure(s) exert the effect on outcome(s). However, existing methods fail to consider the case where (1) both exposures and mediators are potentially high-dimensional and (2) it is very likely that some important confounding variables are unmeasured or latent; both issues are quite common in practice. To the best of our knowledge, however, no methods have been developed to address these challenges with statistical guarantees. ResultsIn this article, we propose a new method for HIgh-dimensional LAtent-confounding Mediation Analysis, abbreviated as "HILAMA", that considers both high-dimensional exposures and mediators, and more importantly, the possible existence of latent confounding variables. HILAMA achieves false discovery rate (FDR) control under finite sample size for multiple mediation effect testing. The proposed method is evaluated through extensive simulation experiments, demonstrating its improved stability in FDR control and superior power in finite sample size compared to existing competitive methods. Furthermore, our method is applied to the proteomics-radiomics data from ADNI, identifying some key proteins and brain regions relating to Alzheimers disease. The results show that HILAMA can effectively control FDR and provide valid statistical inference for high dimensional mediation analysis with latent confounding variables. AvailabilityThe R package HILAMA is publicly available at https://github.com/Cinbo-Wang/HILAMA. Contactcinbo_w@sjtu.edu.cn

bioinformatics↗

bioRxiv2023

Toolkits for detailed and high-throughput interrogation of synapses in C. elegans

Visualizing synaptic connectivity has traditionally relied on time-consuming electron microscopy-based imaging approaches. To scale the analysis of synaptic connectivity, fluorescent protein-based techniques have been established, ranging from the labeling of specific pre- or postsynaptic components of chemical or electrical synapses to transsynaptic proximity labeling technology such as GRASP and iBLINC. In this paper, we describe WormPsyQi, a generalizable image analysis pipeline that automatically quantifies synaptically localized fluorescent signals in a high-throughput and robust manner, with reduced human bias. We also present a resource of 30 transgenic strains that label chemical or electrical synapses throughout the nervous system of the nematode C. elegans, using CLA-1, RAB-3, GRASP (chemical synapses), or innexin (electrical synapse) reporters. We show that WormPsyQi captures synaptic structures in spite of substantial heterogeneity in neurite morphology, fluorescence signal, and imaging parameters. We use these toolkits to quantify multiple obvious and subtle features of synapses - such as number, size, intensity, and spatial distribution of synapses - in datasets spanning various regions of the nervous system, developmental stages, and sexes. Although the pipeline is described in the context of synapses, it may be utilized for other punctate signals, such as fluorescently-tagged neurotransmitter receptors and cell adhesion molecules, as well as proteins in other subcellular contexts. By overcoming constraints on time, sample size, cell morphology, and phenotypic space, this work represents a powerful resource for further analysis of synapse biology in C. elegans.

neuroscience↗

bioRxiv2023

Multi-OMIC analysis of Huntington disease reveals a neuroprotective astrocyte state

Huntington disease (HD) is an incurable neurodegenerative disease characterized by neuronal loss and astrogliosis. One hallmark of HD is the selective neuronal vulnerability of striatal medium spiny neurons. To date, the underlying mechanisms of this selective vulnerability have not been fully defined. Here, we employed a multi-omic approach including single nucleus RNAseq (snRNAseq), bulk RNAseq, lipidomics, HTT gene CAG repeat length measurements, and multiplexed immunofluorescence on post-mortem brain tissue from multiple brain regions of HD and control donors. We defined a signature of genes that is driven by CAG repeat length and found it enriched in astrocytic and microglial genes. Moreover, weighted gene correlation network analysis showed loss of connectivity of astrocytic and microglial modules in HD and identified modules that correlated with CAG-repeat length which further implicated inflammatory pathways and metabolism. We performed lipidomic analysis of HD and control brains and identified several lipid species that correlate with HD grade, including ceramides and very long chain fatty acids. Integration of lipidomics and bulk transcriptomics identified a consensus gene signature that correlates with HD grade and HD lipidomic abnormalities and implicated the unfolded protein response pathway. Because astrocytes are critical for brain lipid metabolism and play important roles in regulating inflammation, we analyzed our snRNAseq dataset with an emphasis on astrocyte pathology. We found two main astrocyte types that spanned multiple brain regions; these types correspond to protoplasmic astrocytes, and fibrous-like - CD44-positive, astrocytes. HD pathology was differentially associated with these cell types in a region-specific manner. One protoplasmic astrocyte cluster showed high expression of metallothionein genes, the depletion of this cluster positively correlated with the depletion of vulnerable medium spiny neurons in the caudate nucleus. We confirmed that metallothioneins were increased in cingulate HD astrocytes but were unchanged or even decreased in caudate astrocytes. We combined existing genome-wide association studies (GWAS) with a GWA study conducted on HD patients from the original Venezuelan cohort and identified a single-nucleotide polymorphism in the metallothionein gene locus associated with delayed age of onset. Functional studies found that metallothionein overexpressing astrocytes are better able to buffer glutamate and were neuroprotective of patient-derived directly reprogrammed HD MSNs as well as against rotenone-induced neuronal death in vitro. Finally, we found that metallothionein-overexpressing astrocytes increased the phagocytic activity of microglia in vitro and increased the expression of genes involved in fatty acid binding. Together, we identified an astrocytic phenotype that is regionally-enriched in less vulnerable brain regions that can be leveraged to protect neurons in HD.

neuroscience↗

bioRxiv2023

PVN microglia via P2Y12 transmit hemodynamic signal to promote sympathetic excitation in hypertension

Hypertension is usually accompanied with an elevated sympathetic tonicity, but how sympathetic hyperactivity is triggered is not fully understood. Recent advances reveal that microglia-centered neuroinflammation contributes to sympathetic excitation in hypertension. In this study, we performed a temporospatial analysis of microglia at both morphological and transcriptomic levels, and found that microglia in the hypothalamic paraventricular nucleus (PVN) were early responders to hypertensive challenges. PVN is the central hub for maintaining cardiovascular function via regulation of fluid balance and sympathetic outflow. Comprehensive vasculature analyses unveiled that PVN was characterized by high capillary density, thin vessel diameter, and complex vascular topology among brain regions. As such, PVN is susceptible to the penetration of ATP released from the vasculature in response to hemodynamic disturbance after blood pressure increase. ATP ligation to microglial P2Y12 receptor is responsible for the microglial accumulation and activation in the PVN. Furthermore, either pharmacological blockade or genetic ablation of microglial P2Y12 could substantially restrain blood pressure increase under hypertensive challenge. Together, these findings disclose that a unique vasculature pattern results in the vulnerability of PVN pre-sympathetic neurons to hypertension-associated insults, which is mediated by microglia.

neuroscience↗

bioRxiv2023

Large-scale population genomics of Malayan pangolins reveals deep diversification and a new species

BackgroundArchipelagos and oceanic islands often present high percentage of endemism due to rapid speciation. The Malayan pangolin is a species distributing at both mainland (southern Yunnan, China) and oceanic islands via Malayan peninsula, which may result in deep differentiation among populations. In-depth investigation of population structure and genetic consequences for such species is of vital importance for their protection and conservation, practically for the critically endangered Malayan pangolin that is suffering from poaching, illegal trade, and habitat loss. ResultsHere we carried out a large-scale population genomic analysis for Malayan pangolins, and revealed three highly distinct genetic populations in this species, two of which are now being reported for the first time. Based on multiple lines of genomic and morphological evidence, we postulate the existence of a new pangolin species (Manis_1). Genetic diversity and recent inbreeding were both at a moderate level for both Malayan pangolins and Manis_1, but mainland Malayan pangolins presented relatively lower genetic diversity, higher inbreeding and fitness cost than island populations. ConclusionsWe found extremely deep and graded differentiation in Malayan pangolins, with two newly discovered genetic populations and a new pangolin species that is closely related to the Philippine pangolin than the typical Malayan pangolin, but a distant relative of the Indian pangolin. Anthropogenic factors did not significantly weaken the basis of genetic sustainability for Malayan pangolins, but mainland Malayan pangolins should be paid more attention for conservation due to higher genetic risks than island populations.

genomics↗

bioRxiv2023

Next-generation anti-PD-L1/IL-15 immunocytokine elicits superior antitumor immunity in cold tumors with minimal toxicity

Immunocytokines, such as anti-PD-L1/IL-15, have shown promising efficacy in preclinical studies, but their clinical development still faces severe safety concerns, with the problem not easily overcome by simply reducing the cytokine activity. We proposed a next-generation immunocytokine concept of designing a tumor-conditional anti-PD-L1/IL-15 prodrug (LH05), which innovatively masks IL-15 with steric hindrance of its flanking moieties of anti-PD-L1 and IL-15R-sushi domain. The design successfully attenuated the cytokine sink effect of IL-15 and resulted in a significantly reduced systemic toxicity when compared to wild-type anti-PD-L1/IL-15. LH05 would be specifically cleaved in the tumor microenvironment (TME) to release the active IL-15/IL-15R-sushi domain (ILR) in a proteolytic cleavage-dependent manner and exhibited potent antitumor effects in mouse syngeneic models. Mechanistically, the antitumor efficacy of LH05 was dependent on both innate and adaptive immunity, which altered the TME to Th1-type by recruiting and stimulating both NK and CD8+ T cells and fired up cold tumors. LH05 also showed superior efficacy in restoring immunotherapy response in a refractory U251 xenograft model. Collectively, we introduced a novel next-generation immunocytokine strategy for tumor immunotherapy, contributing to the establishment of optimal treatment for patients with resistance to immune checkpoint inhibitors or cold tumors.

bioengineering↗

bioRxiv2023

Reprogramming Yarrowia lipolytica metabolism for efficient synthesis of itaconic acid from flask to semi-pilot scale

Itaconic acid is an emerging platform chemical with extensive applications. It is currently produced by Aspergillus terreus through biological fermentation. However, A. terreus is a fungal pathogen and needs additional morphology controls, and therefore the production remains problematic. Here, we reprogrammed the GRAS yeast Yarrowia lipolytica metabolism for competitive itaconic acid production. After redirecting the flux of lipid accumulation as carbon sink, we evaluated itaconic acid production both inside and outside the mitochondria, and fine modulated its synthetic pathway. We then mimicked the regulation of nitrogen limitation in nitrogen replete conditions through down regulation of IDH by weak promoter changing, RNAi, or CRISPRi. Ultimately, we optimized fermentation parameters for fed-batch cultivations, and produced itaconic acid with titres of 130.1 g/L in 1L bioreactors and 94.8 g/L in a 50L bioreactor on semi-pilot scale. Our finds provide effective approaches for harnessing GRAS microorganism for competitive industrial itaconic acid production.

bioengineering↗

bioRxiv2023

Automated cell annotation in multi-cell images using an improved CRF_ID algorithm

Cell identification is an important yet difficult process in data analysis of biological images. Previously, we developed an automated cell identification method called CRF_ID and demonstrated its high performance in C. elegans whole-brain images (Chaudhary et al, 2021). However, because the method was optimized for whole-brain imaging, comparable performance could not be guaranteed for application in commonly used C. elegans multi-cell images that display a subpopulation of cells. Here, we present an advance CRF_ID 2.0 that expands the generalizability of the method to multi-cell imaging beyond whole-brain imaging. To illustrate the application of the advance, we show the characterization of CRF_ID 2.0 in multi-cell imaging and cell-specific gene expression analysis in C. elegans. This work demonstrates that high accuracy automated cell annotation in multi-cell imaging can expedite cell identification and reduce its subjectivity in C. elegans and potentially other biological images of various origins.

bioengineering↗

bioRxiv2023

Transformer with Convolution and Graph-Node co-embedding: An accurate and interpretable vision backbone for predicting gene expressions from local histopathological image

Inferring gene expressions from histopathological images has always been a fascinating but challenging task due to the huge differences between the two modal data. Previous works have used modified DenseNet121 to encode the local images and make gene expression predictions. And later works improved the prediction accuracy of gene expression by incorporating the coordinate information from images and using all spots in the tissue region as input. While these methods were limited in use due to model complexity, large demand on GPU memory, and insufficient encoding of local images, thus the results had low interpretability, relatively low accuracy, and over-smooth prediction of gene expression among neighbor spots. In this paper, we propose TCGN, (Transformer with Convolution and Graph-Node co-embedding method) for gene expression prediction from H&E stained pathological slide images. TCGN consists of convolutional layers, transformer encoders, and graph neural networks, and is the first to integrate these blocks in a general and interpretable computer vision backbone for histopathological image analysis. We trained TCGN and compared its performance with three existing methods on a publicly available spatial transcriptomic dataset. Even in the absence of the coordinates information and neighbor spots, TCGN still outperformed the existing methods by 5% and achieved 10 times higher prediction accuracy than the counterpart model. Besides its higher accuracy, our model is also small enough to be run on a personal computer and does not need complex building graph preprocessing compared to the existing methods. Moreover, TCGN is interpretable in recognizing special cell morphology and cell-cell interactions compared to models using all spots as input that are not interpretable. A more accurate omics information prediction from pathological images not only links genotypes to phenotypes so that we can predict more biomarkers that are expensive to test from histopathological images that are low-cost to obtain, but also provides a theoretical basis for future modeling of multi-modal data. Our results support that TCGN is a useful tool for inferring gene expressions from histopathological images and other potential histopathological image analysis studies. HighlightsO_LIFirst deep learning model to integrate CNN, GNN, and transformer for image analysis C_LIO_LIAn interpretable model that uses cell morphology and organizations to predict genes C_LIO_LIHigher gene expression prediction accuracy without global information C_LIO_LIAccurately predicted genes are related to immune escape and abnormal metabolism C_LIO_LIPredict important biomarkers for breast cancer accurately from cheaper images C_LI Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=123 SRC="FIGDIR/small/542669v1_ufig1.gif" ALT="Figure 1"> View larger version (51K): org.highwire.dtl.DTLVardef@1c6a4dforg.highwire.dtl.DTLVardef@7266d4org.highwire.dtl.DTLVardef@bcfd0aorg.highwire.dtl.DTLVardef@188d452_HPS_FORMAT_FIGEXP M_FIG C_FIG

bioinformatics↗