A spatially resolved single-cell genomic atlas of the adult human breast.

The adult human breast is comprised of an intricate network of epithelial ducts and lobules that are embedded in connective and adipose tissue1-3. Although most previous studies have focused on the breast epithelial system4-6, many of the non-epithelial cell types remain understudied. Here we constructed the comprehensive Human Breast Cell Atlas (HBCA) at single-cell and spatial resolution. Our single-cell transcriptomics study profiled 714,331 cells from 126 women, and 117,346 nuclei from 20 women, identifying 12 major cell types and 58 biological cell states. These data reveal abundant perivascular, endothelial and immune cell populations, and highly diverse luminal epithelial cell states. Spatial mapping using four different technologies revealed an unexpectedly rich ecosystem of tissue-resident immune cells, as well as distinct molecular differences between ductal and lobular regions. Collectively, these data provide a reference of the adult normal breast tissue for studying mammary biology and diseases such as breast cancer.

A spatially resolved single cell genomic atlas of the adult human breast.

The adult human breast comprises an intricate network of epithelial ducts and lobules that are embedded in connective and adipose tissue. While previous studies have mainly focused on the breast epithelial system, many of the non-epithelial cell types remain understudied. Here, we constructed a comprehensive Human Breast Cell Atlas (HBCA) at single-cell and spatial resolution. Our single-cell transcriptomics data profiled 535,941 cells from 62 women, and 120,024 nuclei from 20 women, identifying 11 major cell types and 53 cell states. These data revealed abundant pericyte, endothelial and immune cell populations, and highly diverse luminal epithelial cell states. Our spatial mapping using three technologies revealed an unexpectedly rich ecosystem of tissue-resident immune cells in the ducts and lobules, as well as distinct molecular differences between ductal and lobular regions. Collectively, these data provide an unprecedented reference of adult normal breast tissue for studying mammary biology and disease states such as breast cancer.

Left-Censored Missing Value Imputation Approach for MS-Based Proteomics Data with GSimp.

Missing values caused by the limit of detection or quantification (LOD/LOQ) were widely observed in mass spectrometry (MS)-based omics studies and could be recognized as missing not at random (MNAR). MNAR leads to biased statistical estimations and jeopardizes downstream analyses. Although a wide range of missing value imputation methods was developed for omics studies, a limited number of methods were designed appropriately for the situation of MNAR. To facilitate MS-based omics studies, we introduce GSimp, a Gibbs sampler-based missing value imputation approach, to deal with left-censor missing values in MS-proteomics datasets. In this book, we explain the MNAR and elucidate the usage of GSimp for MNAR in detail.

Spatial charting of single-cell transcriptomes in tissues.

Single-cell RNA sequencing methods can profile the transcriptomes of single cells but cannot preserve spatial information. Conversely, spatial transcriptomics assays can profile spatial regions in tissue sections, but do not have single-cell resolution. Here, we developed a computational method called CellTrek that combines these two datasets to achieve single-cell spatial mapping through coembedding and metric learning approaches. We benchmarked CellTrek using simulation and in situ hybridization datasets, which demonstrated its accuracy and robustness. We then applied CellTrek to existing mouse brain and kidney datasets and showed that CellTrek can detect topological patterns of different cell types and cell states. We performed single-cell RNA sequencing and spatial transcriptomics experiments on two ductal carcinoma in situ tissues and applied CellTrek to identify tumor subclones that were restricted to different ducts, and specific T cell states adjacent to the tumor areas. Our data show that CellTrek can accurately map single cells in diverse tissue types to resolve their spatial organization.

Author Correction: Analysis of the Human Protein Atlas Image Classification competition.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Peripheral serum metabolomic profiles inform central cognitive impairment.

The incidence of Alzheimer's disease (AD) increases with age and is becoming a significant cause of worldwide morbidity and mortality. However, the metabolic perturbation behind the onset of AD remains unclear. In this study, we performed metabolite profiling in both brain (n = 109) and matching serum samples (n = 566) to identify differentially expressed metabolites and metabolic pathways associated with neuropathology and cognitive performance and to identify individuals at high risk of developing cognitive impairment. The abundances of 6 metabolites, glycolithocholate (GLCA), petroselinic acid, linoleic acid, myristic acid, palmitic acid, palmitoleic acid and the deoxycholate/cholate (DCA/CA) ratio, along with the dysregulation scores of 3 metabolic pathways, primary bile acid biosynthesis, fatty acid biosynthesis, and biosynthesis of unsaturated fatty acids showed significant differences across both brain and serum diagnostic groups (P-value < 0.05). Significant associations were observed between the levels of differential metabolites/pathways and cognitive performance, neurofibrillary tangles, and neuritic plaque burden. Metabolites abundances and personalized metabolic pathways scores were used to derive machine learning models, respectively, that could be used to differentiate cognitively impaired persons from those without cognitive impairment (median area under the receiver operating characteristic curve (AUC) = 0.772 for the metabolite level model; median AUC = 0.731 for the pathway level model). Utilizing these two models on the entire baseline control group, we identified those who experienced cognitive decline in the later years (AUC = 0.804, sensitivity = 0.722, specificity = 0.749 for the metabolite level model; AUC = 0.778, sensitivity = 0.633, specificity = 0.825 for the pathway level model) and demonstrated their pre-AD onset prediction potentials. Our study provides a proof-of-concept that it is possible to discriminate antecedent cognitive impairment in older adults before the onset of overt clinical symptoms using metabolomics. Our findings, if validated in future studies, could enable the earlier detection and intervention of cognitive impairment that may halt its progression.

Serum metabolite profiles are associated with the presence of advanced liver fibrosis in Chinese patients with chronic hepatitis B viral infection.

BACKGROUND: Accurate and noninvasive diagnosis and staging of liver fibrosis are essential for effective clinical management of chronic liver disease (CLD). We aimed to identify serum metabolite markers that reliably predict the stage of fibrosis in CLD patients. METHODS: We quantitatively profiled serum metabolites of participants in 2 independent cohorts. Based on the metabolomics data from cohort 1 (504 HBV associated liver fibrosis patients and 502 normal controls, NC), we selected a panel of 4 predictive metabolite markers. Consequently, we constructed 3 machine learning models with the 4 metabolite markers using random forest (RF), to differentiate CLD patients from normal controls (NC), to differentiate cirrhosis patients from fibrosis patients, and to differentiate advanced fibrosis from early fibrosis, respectively. RESULTS: The panel of 4 metabolite markers consisted of taurocholate, tyrosine, valine, and linoelaidic acid. The RF models of the metabolite panel demonstrated the strongest stratification ability in cohort 1 to diagnose CLD patients from NC (area under the receiver operating characteristic curve (AUROC) = 0.997 and the precision-recall curve (AUPR) = 0.994), to differentiate fibrosis from cirrhosis (0.941, 0.870), and to stage liver fibrosis (0.918, 0.892). The diagnostic accuracy of the models was further validated in an independent cohort 2 consisting of 300 CLD patients with chronic HBV infection and 90 NC. The AUCs of the models were consistently higher than APRI, FIB-4, and AST/ALT ratio, with both greater sensitivity and specificity. CONCLUSIONS: Our study showed that this 4-metabolite panel has potential usefulness in clinical assessments of CLD progression in patients with chronic hepatitis B virus infection.

Publisher Correction: Analysis of the Human Protein Atlas Image Classification competition.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Publisher Correction: Analysis of the Human Protein Atlas Image Classification competition.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Theabrownin from Pu-erh tea attenuates hypercholesterolemia via modulation of gut microbiota and bile acid metabolism.

Pu-erh tea displays cholesterol-lowering properties, but the underlying mechanism has not been elucidated. Theabrownin is one of the most active and abundant pigments in Pu-erh tea. Here, we show that theabrownin alters the gut microbiota in mice and humans, predominantly suppressing microbes associated with bile-salt hydrolase (BSH) activity. Theabrownin increases the levels of ileal conjugated bile acids (BAs) which, in turn, inhibit the intestinal FXR-FGF15 signaling pathway, resulting in increased hepatic production and fecal excretion of BAs, reduced hepatic cholesterol, and decreased lipogenesis. The inhibition of intestinal FXR-FGF15 signaling is accompanied by increased gene expression of enzymes in the alternative BA synthetic pathway, production of hepatic chenodeoxycholic acid, activation of hepatic FXR, and hepatic lipolysis. Our results shed light into the mechanisms behind the cholesterol- and lipid-lowering effects of Pu-erh tea, and suggest that decreased intestinal BSH microbes and/or decreased FXR-FGF15 signaling may be potential anti-hypercholesterolemia and anti-hyperlipidemia therapies.

Analysis of the Human Protein Atlas Image Classification competition.

Pinpointing subcellular protein localizations from microscopy images is easy to the trained eye, but challenging to automate. Based on the Human Protein Atlas image collection, we held a competition to identify deep learning solutions to solve this task. Challenges included training on highly imbalanced classes and predicting multiple labels per image. Over 3 months, 2,172 teams participated. Despite convergence on popular networks and training techniques, there was considerable variety among the solutions. Participants applied strategies for modifying neural networks and loss functions, augmenting data and using pretrained networks. The winning models far outperformed our previous effort at multi-label classification of protein localization patterns by ~20%. These models can be used as classifiers to annotate new images, feature extractors to measure pattern similarity or pretrained networks for a wide range of biological applications.

Strategy for Intercorrelation Identification between Metabolome and Microbiome.

Accumulating evidence points to the strong and complicated associations between the metabolome and the microbiome, which play diverse roles in physiology and pathology. Various correlation analysis approaches were applied to identify microbe-metabolite associations. Given the strengths and weaknesses of the existing methods and considering the characteristics of different types of omics data, we designed a special strategy, called Generalized coRrelation analysis for Metabolome and Microbiome (GRaMM), for the intercorrelation discovery between the metabolome and microbiome. GRaMM can properly deal with two types of omics data, the effect of confounders, and both linear and nonlinear correlations by integrating several complementary methods such as the classical linear regression, the emerging maximum information coefficient (MIC), the metabolic confounding effect elimination (MCEE), and the centered log-ratio transformation (CLR). GRaMM contains four sequential computational steps: (1) metabolic and microbial data preprocessing, (2) linear/nonlinear type identification, (3) data correction and correlation detection, and (4) p value correction. The performances of GRaMM, including the accuracy, sensitivity, specificity, false positive rate, applicability, and effects of preprocessing and confounder adjustment steps, were evaluated and compared with three other methods in multiple simulated and real-world datasets. To our knowledge, GRaMM is the first strategy designed for the intercorrelation analysis between metabolites and microbes. The Matlab function and an R package were developed and are freely available for academic use (comply with GNU GPL.V3 license).

Transcriptomics Associates Molecular Features with 18F-Fluorocholine PET/CT Imaging Phenotype and Its Potential Relationship to Survival in Hepatocellular Carcinoma.

Studies involving transcriptomics have revealed multiple molecular subtypes of hepatocellular carcinoma (HCC). Positron emission tomography/computed tomography (PET/CT) has also identified distinct molecular imaging subtypes, including those with increased and decreased choline metabolism as measured by the tissue uptake of the radiopharmaceutical 18F-fluorocholine. Gene signatures reflecting the molecular heterogeneity of HCC may identify the biological and clinical significance of these imaging subtypes. In this study, 41 patients underwent 18F-fluorocholine PET/CT, followed by tumor resection and gene expression profiling. Over- and underexpressed components of previously published gene signatures were evaluated for enrichment between tumors with high and low 18F-fluorocholine uptake using gene set analysis. Significant gene sets were enumerated by FDR based on phenotype permutation. Associations with overall survival were analyzed by univariate and multivariate proportional hazards regression. Ten gene sets related to HCC were significantly associated with high tumor 18F-fluorocholine uptake at FDR q < 0.05, including those from three different clinical molecular classification systems and two prognostic signatures for HCC that showed predictive value in the study cohort. Tumor avidity for 18F-fluorocholine was associated with favorable characteristics based on these signatures with lower mortality based on survival analysis (HR 0.36; 95% confidence interval, 0.14-0.95). Tumors demonstrating high 18F-fluorocholine uptake were also enriched for genes involved in oxidative phosphorylation, fatty acid metabolism, peroxisome, bile acid metabolism, xenobiotic metabolism, and adipogenesis. These results provide a pathobiological framework to further evaluate 18F-fluorocholine PET/CT as a molecular and prognostic classifier in HCC. SIGNIFICANCE: A pathobiological framework for HCC brings together multiple prognostically relevant gene signatures via convergence with 18F-fluorocholine PET/CT imaging phenotype.

Evaluation of metabolite-microbe correlation detection methods.

Different correlation detection methods have been specifically designed for the microbiome data analysis considering the compositional data structure and different sequencing depths. Along with the speedy development of omics studies, there is an increasing interest in discovering the biological associations between microbes and host metabolites. This raises the need of finding proper statistical methods that facilitate the correlation analysis across different omics studies. Here, we comprehensively evaluated six different correlation methods, i.e., Pearson correlation, Spearman correlation, Sparse Correlations for Compositional data (SparCC), Correlation inference for Compositional data through Lasso (CCLasso), Mutual Information Coefficient (MIC), and Cosine similarity methods, for the correlations detection between microbes and metabolites. Three simulated and two real-world data sets (from public databases and our lab) were used to examine the performance of each method regarding its specificity, sensitivity, similarity, accuracy, and stability with different sparsity. Our results indicate that although each method has its own pros and cons in different scenarios, Spearman correlation and MIC outperform the others with their overall performances. A strategic guidance was also proposed for the correlation analysis between microbe and metabolite.

Modification effect of fenofibrate therapy, a longitudinal epigenomic-wide methylation study of triglycerides levels in the GOLDN study.

BACKGROUND: Identification of interactions between epigenetic factors and treatments might lead to personalized intervention of diseases. This paper aims to examine the modification effect of fenofibrate therapy on the association of methylation levels and fasting blood triglycerides (TG), and the related biological pathways among methylation sites. RESULTS: Mixed-effects models were employed to assess pre- and posttreatment associations and drug modification effects simultaneously. Five cytosine-phosphate-guanine (CpG) sites were found to be associated with TG levels before and after the fenofibrate therapy: cg00574958, cg17058475, and cg01082498 on CPT1A gene, chromosome 11; cg03725309 on SARS, chromosome 1; and cg06500161 on ABCG1, chromosome 21. In addition, fenofibrate therapy modified the methylation levels on the following 4 CpG sites: cg20015535 (gene EGLN1, chromosome 1); cg24870738 (gene RNF220, chromosome 1); cg06891775 (gene LOC283050, chromosome 10); and cg00607630 (gene USP7, chromosome 16). Further, gene set enrichment analysis (GSEA) identified cancer- and metabolism-related pathways that were associated with TG-related CpG sites. CONCLUSIONS: We identified modification effects of fenofibrate on the associations between blood TG levels and several CpG sites. Pathway enrichment analysis indicated the alternations in some metabolism and cancer-related pathways. Our findings have important implications for future research in pharmacoepigenetics and personalized medicine.

Metabotypes Related to Meat and Vegetable Intake Reflect Microbial, Lipid and Amino Acid Metabolism in Healthy People.

SCOPE: The objective of this study is to develop a new methodology to identify the relationship between dietary patterns and metabolites indicative of food intake and metabolism. METHODS AND RESULTS: Plasma and urine samples from healthy Swiss subjects (n = 89) collected over two time points are analyzed for a panel of host-microbial metabolites using GC- and LC-MS. Dietary intake is evaluated using a validated food frequency questionnaire. Dietary pattern clusters and relationships with metabolites are determined using Non-Negative Matrix Factorization (NNMF) and Sparse Generalized Canonical Correlation Analysis (SGCCA). Use of NNMF allows detection of latent diet clusters in this population, which describes a high intake of meat or vegetables. SGCCA associates these clusters to i) diet-host microbial and lipid associated bile acid metabolism, and ii) essential amino acid metabolism. CONCLUSION: This novel application of NNMF and SGCCA allows detection of distinct metabotypes for meat and vegetable dietary patterns in a heterogeneous population. As many of the metabolites associated with meat or vegetable intake are the result of host-microbiota interactions, the findings support a role for microbiota mediating the metabolic imprinting of different dietary choices.

Clinical prediction of HBV and HCV related hepatic fibrosis using machine learning.

Clinical prediction of advanced hepatic fibrosis (HF) and cirrhosis has long been challenging due to the gold standard, liver biopsy, being an invasive approach with certain limitations. Less invasive blood test tandem with a cutting-edge machine learning algorithm shows promising diagnostic potential. In this study, we constructed and compared machine learning methods with the FIB-4 score in a discovery dataset (n = 490) of hepatitis B virus (HBV) patients. Models were validated in an independent HBV dataset (n = 86). We further employed these models on two independent hepatitis C virus (HCV) datasets (n = 254 and 230) to examine their applicability. In the discovery data, gradient boosting (GB) stably outperformed other methods as well as FIB-4 scores (p < .001) in the prediction of advanced HF and cirrhosis. In the HBV validation dataset, for classification between early and advanced HF, the area under receiver operating characteristic curves (AUROC) of GB model was 0.918, while FIB-4 was 0.841; for classification between non-cirrhosis and cirrhosis, GB showed AUROC of 0.871, while FIB-4 was 0.830. Additionally, GB-based prediction demonstrated good classification capacity on two HCV datasets while higher cutoffs for both GB and FIB-4 scores were required to achieve comparable specificity and sensitivity. Using the same parameters as FIB-4, the GB-based prediction system demonstrated steady improvements relative to FIB-4 in HBV and HCV cohorts with different cutoff values required in different etiological groups. A user-friendly web tool, LiveBoost, makes our prediction models freely accessible for further clinical studies and applications.

Metabolome and gut microbiota variation with long-term intake of Panax ginseng extracts on rats.

Ginseng, a widely used functional food and food additive, has been proven to have promotion effects of health on the body. However, whether the long-term intake of Ginseng is beneficial or has side effects on an organism is still unclear. In this study, untargeted GC-TOFMS metabolomic analysis of serum, cecum and ileum intestinal contents was conducted to understand the effect of the long-term intake of Ginseng extracts. 16S rRNA microbial sequencing technology was applied to investigate the effect of Ginseng extracts on the structure of gut microbiota. Cytokines in spleen were detected to determine the effect of Ginseng extracts on the immune system. Compared to control groups, the metabolites in serum, cecum and ileum, such as amino acids, amines and other metabolites related to carbohydrate metabolism, significantly varied between the C and GS groups. Ginseng extracts affected the structure of gut microbiota with a decreased abundance of TM7, while the abundance of Proteobacteria, Methylobacteriaceae, Parasutterella, Sutterella increased in the GS group. The increased abundance of Bifidobacterium and Lactobacillus demonstrated that Ginseng extracts contribute to probiotic amplification. Highly correlated with Bifidobacterium and Lactobacillus, interleukin 4 (IL4), IL10 and immunoglobulin A (IgA) levels were significantly elevated after the long-term intake of Ginseng extracts. These results indicated that the long-term administration of Ginseng extracts positively affected the host-gut metabolism, immune system, the anti-inflammation process and the gut intestinal microbiota structure.

GSimp: A Gibbs sampler based left-censored missing value imputation approach for metabolomics studies.

Left-censored missing values commonly exist in targeted metabolomics datasets and can be considered as missing not at random (MNAR). Improper data processing procedures for missing values will cause adverse impacts on subsequent statistical analyses. However, few imputation methods have been developed and applied to the situation of MNAR in the field of metabolomics. Thus, a practical left-censored missing value imputation method is urgently needed. We developed an iterative Gibbs sampler based left-censored missing value imputation approach (GSimp). We compared GSimp with other three imputation methods on two real-world targeted metabolomics datasets and one simulation dataset using our imputation evaluation pipeline. The results show that GSimp outperforms other imputation methods in terms of imputation accuracy, observation distribution, univariate and multivariate analyses, and statistical sensitivity. Additionally, a parallel version of GSimp was developed for dealing with large scale metabolomics datasets. The R code for GSimp, evaluation pipeline, tutorial, real-world and simulated targeted metabolomics datasets are available at: https://github.com/WandeRum/GSimp.

Missing Value Imputation Approach for Mass Spectrometry-based Metabolomics Data.

Missing values exist widely in mass-spectrometry (MS) based metabolomics data. Various methods have been applied for handling missing values, but the selection can significantly affect following data analyses. Typically, there are three types of missing values, missing not at random (MNAR), missing at random (MAR), and missing completely at random (MCAR). Our study comprehensively compared eight imputation methods (zero, half minimum (HM), mean, median, random forest (RF), singular value decomposition (SVD), k-nearest neighbors (kNN), and quantile regression imputation of left-censored data (QRILC)) for different types of missing values using four metabolomics datasets. Normalized root mean squared error (NRMSE) and NRMSE-based sum of ranks (SOR) were applied to evaluate imputation accuracy. Principal component analysis (PCA)/partial least squares (PLS)-Procrustes analysis were used to evaluate the overall sample distribution. Student's t-test followed by correlation analysis was conducted to evaluate the effects on univariate statistics. Our findings demonstrated that RF performed the best for MCAR/MAR and QRILC was the favored one for left-censored MNAR. Finally, we proposed a comprehensive strategy and developed a public-accessible web-tool for the application of missing value imputation in metabolomics ( https://metabolomics.cc.hawaii.edu/software/MetImp/ ).