CRISP: a deep learning architecture for GC × GC-TOFMS contour ROI identification, simulation and analysis in imaging metabolomics.

Two-dimensional gas chromatography-time-of-flight mass spectrometry (GC × GC-TOFMS) provides a large amount of molecular information from biological samples. However, the lack of a comprehensive compound library or customizable bioinformatics tool is currently a challenge in GC × GC-TOFMS data analysis. We present an open-source deep learning (DL) software called contour regions of interest (ROI) identification, simulation and untargeted metabolomics profiler (CRISP). CRISP integrates multiple customizable deep neural network architectures for assisting the semi-automated identification of ROIs, contour synthesis, resolution enhancement and classification of GC × GC-TOFMS-based contour images. The approach includes the novel aggregate feature representative contour (AFRC) construction and stacked ROIs. This generates an unbiased contour image dataset that enhances the contrasting characteristics between different test groups and can be suitable for small sample sizes. The utility of the generative models and the accuracy and efficacy of the platform were demonstrated using a dataset of GC × GC-TOFMS contour images from patients with late-stage diabetic nephropathy and healthy control groups. CRISP successfully constructed AFRC images and identified over five ROIs to create a deepstacked dataset. The high fidelity, 512 × 512-pixels generative model was trained as a generator with a Fréchet inception distance of <47.00. The trained classifier achieved an AUROC of >0.96 and a classification accuracy of >95.00% for datasets with and without column bleed. Overall, CRISP demonstrates good potential as a DL-based approach for the rapid analysis of 4-D GC × GC-TOFMS untargeted metabolite profiles by directly implementing contour images. CRISP is available at https://github.com/vivekmathema/GCxGC-CRISP.

GC × GC-TOFMS metabolomics analysis identifies elevated levels of plasma sugars and sugar alcohols in diabetic mellitus patients with kidney failure.

Two dimensional GC (GC × GC)-time-of-flight mass spectrometry (TOFMS) has been used to improve accurate metabolite identification in the chemical industry, but this method has not been applied as readily in biomedical research. Here, we evaluated and validated the performance of high resolution GC × GC-TOFMS against that of GC-TOFMS for metabolomics analysis of two different plasma matrices, from healthy controls (CON) and diabetes mellitus (DM) patients with kidney failure (DM with KF). We found GC × GC-TOFMS outperformed traditional GC-TOFMS in terms of separation performance and metabolite coverage. Several metabolites from both the CON and DM with KF matrices, such as carbohydrates and carbohydrate-conjugate metabolites, were exclusively detected using GC × GC-TOFMS. Additionally, we applied this method to characterize significant metabolites in the DM with KF group, with focused analysis of four metabolite groups: sugars, sugar alcohols, amino acids, and free fatty acids. Our plasma metabolomics results revealed 35 significant metabolites (12 unique and 23 concentration-dependent metabolites) in the DM with KF group, as compared with those in the CON and DM groups (N = 20 for each group). Interestingly, we determined 17 of the 35 (14/17 verified with reference standards) significant metabolites identified from both the analyses were metabolites from the sugar and sugar alcohol groups, with significantly higher concentrations in the DM with KF group than in the CON and DM groups. Enrichment analysis of these 14 metabolites also revealed that alterations in galactose metabolism and the polyol pathway are related to DM with KF. Overall, our application of GC × GC-TOFMS identified key metabolites in complex plasma matrices.

Revisiting chloroplast genomic landscape and annotation towards comparative chloroplast genomes of Rhamnaceae.

BACKGROUND: Massive parallel sequencing technologies have enabled the elucidation of plant phylogenetic relationships from chloroplast genomes at a high pace. These include members of the family Rhamnaceae. The current Rhamnaceae phylogenetic tree is from 13 out of 24 Rhamnaceae chloroplast genomes, and only one chloroplast genome of the genus Ventilago is available. Hence, the phylogenetic relationships in Rhamnaceae remain incomplete, and more representative species are needed. RESULTS: The complete chloroplast genome of Ventilago harmandiana Pierre was outlined using a hybrid assembly of long- and short-read technologies. The accuracy and validity of the final genome were confirmed with PCR amplifications and investigation of coverage depth. Sanger sequencing was used to correct for differences in lengths and nucleotide bases between inverted repeats because of the homopolymers. The phylogenetic trees reconstructed using prevalent methods for phylogenetic inference were topologically similar. The clustering based on codon usage was congruent with the molecular phylogenetic tree. The groups of genera in each tribe were in accordance with tribal classification based on molecular markers. We resolved the phylogenetic relationships among six Hovenia species, three Rhamnus species, and two Ventilago species. Our reconstructed tree provides the most complete and reliable low-level taxonomy to date for the family Rhamnaceae. Similar to other higher plants, the RNA editing mostly resulted in converting serine to leucine. Besides, most genes were subjected to purifying selection. Annotation anomalies, including indel calling errors, unaligned open reading frames of the same gene, inconsistent prediction of intergenic regions, and misannotated genes, were identified in the published chloroplast genomes used in this study. These could be a result of the usual imperfections in computational tools, and/or existing errors in reference genomes. Importantly, these are points of concern with regards to utilizing published chloroplast genomes for comparative genomic analysis. CONCLUSIONS: In summary, we successfully demonstrated the use of comprehensive genomic data, including DNA and amino acid sequences, to build a reliable and high-resolution phylogenetic tree for the family Rhamnaceae. Additionally, our study indicates that the revision of genome annotation before comparative genomic analyses is necessary to prevent the propagation of errors and complications in downstream analysis and interpretation.

A double-blinded, randomized, split-side, vehicle-controlled study of the efficacy of cleanser containing Acanthus ebracteatus Vahl., Suregada multiflora, and Acacia concinna in patients with atopic dermatitis: A pilot study.

BACKGROUND: Barrier repair therapy is the key management approach for both eczematous and non-lesional skin of atopic dermatitis. The use of appropriate cleansers to enhance skin hydration is an adjunctive treatment that increases topical drug penetration. Anti-inflammatory properties of various medicinal plants in tropical Asia have been reported. OBJECTIVE: Investigate the efficacy of herbal cleanser containing a combination of herbal extracts from Acanthus ebracteatus Vahl., Suregada multiflora, and Acacia concinna on seemingly intact skin in patients with atopic dermatitis by measuring improvements in the skin barrier function. METHODS: This 2-week pilot study was a split-side, randomized, double-blinded, vehicle-controlled trial. All patients (n = 30) were asked to use both a cleanser with an active formulation containing the herbal extracts and a vehicle- controlled cleanser on each side of mid-volar forearm. Biophysical assessments including transepidermal water loss (TEWL), skin hydration, skin pH, and skin roughness were performed at baseline and upon study completion. RESULTS: Compared to baseline, the median percentage change in TEWL at the end of the study was significantly greater for the active side 10.4 (-19, 20.7) g/m2h than the control side -13.2 (-28.7, 9.1) g/m2h; p = 0.01. The median percentage change of skin hydration, skin pH, and skin roughness of the active side compared to the control side had no a statistical significance. CONCLUSIONS: This cleanser is beneficial when used as adjunctive therapy. Further studies should evaluate its anti- sinflammatory properties in the remedy or active phase of atopic dermatitis or other inflammatory skin diseases.

Integrated Metabolomics and Proteomics Highlight Altered Nicotinamide- and Polyamine Pathways in Lung Adenocarcinoma.

Lung cancer is the leading cause of cancer mortality in the United States with non-small cell lung cancer (NSCLC) adenocarcinoma being the most common histological type. Early perturbations in cellular metabolism are a hallmark of cancer, but the extent of these changes in early stage lung adenocarcinoma remains largely unknown. In the current study, an integrated metabolomics and proteomics approach was utilized to characterize the biochemical and molecular alterations between malignant and matched control tissue from 27 subjects diagnosed with early stage lung adenocarcinoma. Differential analysis identified 71 metabolites and 1102 proteins that delineated tumor from control tissue. Integrated results indicated four major metabolic changes in early stage adenocarcinoma: (1) increased glycosylation and glutaminolysis; (2) elevated Nrf2 activation; (3) increase in nicotinic and nicotinamide salvaging pathways; and (4) elevated polyamine biosynthesis linked to differential regulation of the SAM/nicotinamide methyl-donor pathway. Genomic data from publicly available databases were included to strengthen proteomic findings. Our findings provide insight into the biochemical and molecular biological reprogramming that may accompanies early stage lung tumorigenesis and highlight potential therapeutic targets.

MetaMapR: pathway independent metabolomic network analysis incorporating unknowns.

UNLABELLED: Metabolic network mapping is a widely used approach for integration of metabolomic experimental results with biological domain knowledge. However, current approaches can be limited by biochemical domain or pathway knowledge which results in sparse disconnected graphs for real world metabolomic experiments. MetaMapR integrates enzymatic transformations with metabolite structural similarity, mass spectral similarity and empirical associations to generate richly connected metabolic networks. This open source, web-based or desktop software, written in the R programming language, leverages KEGG and PubChem databases to derive associations between metabolites even in cases where biochemical domain or molecular annotations are unknown. Network calculation is enhanced through an interface to the Chemical Translation System, which allows metabolite identifier translation between >200 common biochemical databases. Analysis results are presented as interactive visualizations or can be exported as high-quality graphics and numerical tables which can be imported into common network analysis and visualization tools. AVAILABILITY AND IMPLEMENTATION: Freely available at http://dgrapov.github.io/MetaMapR/. Requires R and a modern web browser. Installation instructions, tutorials and application examples are available at http://dgrapov.github.io/MetaMapR/. CONTACT: ofiehn@ucdavis.edu.

Data processing solutions to render metabolomics more quantitative: case studies in food and clinical metabolomics using Metabox 2.0.

In classic semiquantitative metabolomics, metabolite intensities are affected by biological factors and other unwanted variations. A systematic evaluation of the data processing methods is crucial to identify adequate processing procedures for a given experimental setup. Current comparative studies are mostly focused on peak area data but not on absolute concentrations. In this study, we evaluated data processing methods to produce outputs that were most similar to the corresponding absolute quantified data. We examined the data distribution characteristics, fold difference patterns between 2 metabolites, and sample variance. We used 2 metabolomic datasets from a retail milk study and a lupus nephritis cohort as test cases. When studying the impact of data normalization, transformation, scaling, and combinations of these methods, we found that the cross-contribution compensating multiple standard normalization (ccmn) method, followed by square root data transformation, was most appropriate for a well-controlled study such as the milk study dataset. Regarding the lupus nephritis cohort study, only ccmn normalization could slightly improve the data quality of the noisy cohort. Since the assessment accounted for the resemblance between processed data and the corresponding absolute quantified data, our results denote a helpful guideline for processing metabolomic datasets within a similar context (food and clinical metabolomics). Finally, we introduce Metabox 2.0, which enables thorough analysis of metabolomic data, including data processing, biomarker analysis, integrative analysis, and data interpretation. It was successfully used to process and analyze the data in this study. An online web version is available at http://metsysbio.com/metabox.

Quantifying fecal and plasma short-chain fatty acids in healthy Thai individuals.

Short-chain fatty acids (SCFAs) are involved in important physiological processes such as gut health and immune response, and changes in SCFA levels can be indicative of disease. Despite the importance of SCFAs in human health and disease, reference values for fecal and plasma SCFA concentrations in healthy individuals are scarce. To address this gap in current knowledge, we developed a simple and reliable derivatization-free GC-TOFMS method for quantifying fecal and plasma SCFAs in healthy individuals. We targeted six linear- and seven branched-SCFAs, obtaining method recoveries of 73-88% and 83-134% in fecal and plasma matrices, respectively. The developed methods are simpler, faster, and more sensitive than previously published methods and are well suited for large-scale studies. Analysis of samples from 157 medically confirmed healthy individuals showed that the total SCFAs in the feces and plasma were 34.1 ± 15.3 µmol/g and 60.0 ± 45.9 µM, respectively. In fecal samples, acetic acid (Ace), propionic acid (Pro), and butanoic acid (But) were all significant, collectively accounting for 89% of the total SCFAs, whereas the only major SCFA in plasma samples was Ace, constituting of 93% of the total plasma SCFAs. There were no statistically significant differences in the total fecal and plasma SCFA concentrations between sexes or among age groups. The data revealed, however, a positive correlation for several nutrients, such as carbohydrate, fat, iron from vegetables, and water, to most of the targeted SCFAs. This is the first large-scale study to report SCFA reference intervals in the plasma and feces of healthy individuals, and thereby delivers valuable data for microbiome, metabolomics, and biomarker research.

Measurement of very low-molecular weight metabolites by traveling wave ion mobility and its use in human urine samples.

The collision cross-sections (CCS) measurement using ion mobility spectrometry (IMS) in combination with mass spectrometry (MS) offers a great opportunity to increase confidence in metabolite identification. However, owing to the lack of sensitivity and resolution, IMS has an analytical challenge in studying the CCS values of very low-molecular-weight metabolites (VLMs ≤ 250 Da). Here, we describe an analytical method using ultrahigh-performance liquid chromatography (UPLC) coupled to a traveling wave ion mobility-quadrupole-time-of-flight mass spectrometer optimized for the measurement of VLMs in human urine samples. The experimental CCS values, along with mass spectral properties, were reported for the 174 metabolites. The experimental data included the mass-to-charge ratio (m/z), retention time (RT), tandem MS (MS/MS) spectra, and CCS values. Among the studied metabolites, 263 traveling wave ion mobility spectrometry (TWIMS)-derived CCS values (TWCCSN2) were reported for the first time, and more than 70% of these were CCS values of VLMs. The TWCCSN2 values were highly repeatable, with inter-day variations of <1% relative standard deviation (RSD). The developed method revealed excellent TWCCSN2 accuracy with a CCS difference (ΔCCS) within ±2% of the reported drift tube IMS (DTIMS) and TWIMS CCS values. The complexity of the urine matrix did not affect the precision of the method, as evidenced by ΔCCS within ±1.92%. According to the Metabolomics Standards Initiative, 55 urinary metabolites were identified with a confidence level of 1. Among these 55 metabolites, 53 (96%) were VLMs. The larger number of confirmed compounds found in this study was a result of the addition of TWCCSN2 values, which clearly increased metabolite identification confidence.

Systems biology of yeast cell death.

Programmed cell death (PCD) (including apoptosis) is an essential process, and many human diseases of high prevalence such as neurodegenerative diseases and cancer are associated with deregulations in the cell death pathways. Yeast Saccharomyces cerevisiae, a unicellular eukaryotic organism, shares with multicellular organisms (including humans) key components and regulators of the PCD machinery. In this article, we review the current state of knowledge about cell death networks, including the modeling approaches and experimental strategies commonly used to study yeast cell death. We argue that the systems biology approach will bring valuable contributions to our understanding of regulations and mechanisms of the complex cell death pathways.

Predicting lupus membranous nephritis using reduced picolinic acid to tryptophan ratio as a urinary biomarker.

The current gold standard for classifying lupus nephritis (LN) progression is a renal biopsy, which is an invasive procedure. Undergoing a series of biopsies for monitoring disease progression and treatments is unlikely suitable for patients with LN. Thus, there is an urgent need for non-invasive alternative biomarkers that can facilitate LN class diagnosis. Such biomarkers will be very useful in guiding intervention strategies to mitigate or treat patients with LN. Urine samples were collected from two independent cohorts. Patients with LN were classified into proliferative (class III/IV) and membranous (class V) by kidney histopathology. Metabolomics was performed to identify potential metabolites, which could be specific for the classification of membranous LN. The ratio of picolinic acid (Pic) to tryptophan (Trp) ([Pic/Trp] ratio) was found to be a promising candidate for LN diagnostic and membranous classification. It has high potential as an alternative biomarker for the non-invasive diagnosis of LN.

ThRSDB: a database of Thai rice starch composition, molecular structure and functionality.

As starch properties can affect end product quality in many ways, rice starch from Thai domesticated cultivars and landraces has been the focus of increasing research interest. Increasing knowledge in this area creates a high demand from the research community for better organized information. The Thai Rice Starch Database (ThRSDB) is an online database containing data extensively curated from original research articles on Thai rice starch composition, molecular structure and functionality. The key aim of the ThRSDB is to facilitate accessibility to dispersed rice starch information for, but not limited to, both research and industrial users. Currently, 373 samples from 191 different Thai rice cultivars have been collected from 39 published articles. The ThRSDB includes the search functions necessary for accessing data together with a user-friendly web interface and interactive visualization tools. We have also demonstrated how the collected data can be efficiently used to observe the relationships between starch parameters and rice cultivars through correlation analysis and Partial Least Squares Discriminant Analysis. Database URL: http://thairicestarch.kku.ac.th.

Deciphering rice metabolic flux reprograming under salinity stress via in silico metabolic modeling.

Rice is one of the most economically important commodities globally. However, rice plants are salt susceptible species in which high salinity can significantly constrain its productivity. Several physiological parameters in adaptation to salt stress have been observed, though changes in metabolic aspects remain to be elucidated. In this study, rice metabolic activities of salt-stressed flag leaf were systematically characterized. Transcriptomics and metabolomics data were combined to identify disturbed pathways, altered metabolites and metabolic hotspots within the rice metabolic network under salt stress condition. Besides, the feasible flux solutions in different context-specific metabolic networks were estimated and compared. Our findings highlighted metabolic reprogramming in primary metabolic pathways, cellular respiration, antioxidant biosynthetic pathways, and phytohormone biosynthetic pathways. Photosynthesis and hexose utilization were among the major disturbed pathways in the stressed flag leaf. Notably, the increased flux distribution of the photorespiratory pathway could contribute to cellular redox control. Predicted flux statuses in several pathways were consistent with the results from transcriptomics, end-point metabolomics, and physiological studies. Our study illustrated that the contextualized genome-scale model together with multi-omics analysis is a powerful approach to unravel the metabolic responses of rice to salinity stress.

Deep metabolome: Applications of deep learning in metabolomics.

In the past few years, deep learning has been successfully applied to various omics data. However, the applications of deep learning in metabolomics are still relatively low compared to others omics. Currently, data pre-processing using convolutional neural network architecture appears to benefit the most from deep learning. Compound/structure identification and quantification using artificial neural network/deep learning performed relatively better than traditional machine learning techniques, whereas only marginally better results are observed in biological interpretations. Before deep learning can be effectively applied to metabolomics, several challenges should be addressed, including metabolome-specific deep learning architectures, dimensionality problems, and model evaluation regimes.

Accounting for biological variation with linear mixed-effects modelling improves the quality of clinical metabolomics data.

Metabolite profiles from biological samples suffer from both technical variations and subject-specific variants. To improve the quality of metabolomics data, conventional data processing methods can be employed to remove technical variations. These methods do not consider sources of subject variation as separate factors from biological factors of interest. This can be a significant issue when performing quantitative metabolomics in clinical trials or screening for a potential biomarker in early-stage disease, because changes in metabolism or a desired-metabolite signal are small compared to the total metabolite signals. As a result, inter-individual variability can interfere subsequent statistical analyses. Here, we propose an additional data processing step using linear mixed-effects modelling to readjust an individual metabolite signal prior to multivariate analyses. Published clinical metabolomics data was used to demonstrate and evaluate the proposed method. We observed a substantial reduction in variation of each metabolite signal after model fitting. A comparison with other strategies showed that our proposed method contributed to improved classification accuracy, precision, sensitivity and specificity. Moreover, we highlight the importance of patient metadata as it contains rich information of subject characteristics, which can be used to model and normalize metabolite abundances. The proposed method is available as an R package lmm2met.

Mitochondrial oxidative stress-induced transcript variants of ATF3 mediate lipotoxic brain microvascular injury.

Elevation of blood triglycerides, primarily triglyceride-rich lipoproteins (TGRL), is an independent risk factor for cardiovascular disease and vascular dementia (VaD). Accumulating evidence indicates that both atherosclerosis and VaD are linked to vascular inflammation. However, the role of TGRL in vascular inflammation, which increases risk for VaD, remains largely unknown and its underlying mechanisms are still unclear. We strived to determine the effects of postprandial TGRL exposure on brain microvascular endothelial cells, the potential risk factor of vascular inflammation, resulting in VaD. We showed in Aung et al., J Lipid Res., 2016 that postprandial TGRL lipolysis products (TL) activate mitochondrial reactive oxygen species (ROS) and increase the expression of the stress-responsive protein, activating transcription factor 3 (ATF3), which injures human brain microvascular endothelial cells (HBMECs) in vitro. In this study, we deployed high-throughput sequencing (HTS)-based RNA sequencing methods and mito stress and glycolytic rate assays with an Agilent Seahorse XF analyzer and profiled the differential expression of transcripts, constructed signaling pathways, and measured mitochondrial respiration, ATP production, proton leak, and glycolysis of HBMECs treated with TL. Conclusions: TL potentiate ROS by mitochondria which activate mitochondrial oxidative stress, decrease ATP production, increase mitochondrial proton leak and glycolysis rate, and mitochondria DNA damage. Additionally, CPT1A1 siRNA knockdown suppresses oxidative stress and prevents mitochondrial dysfunction and vascular inflammation in TL treated HBMECs. TL activates ATF3-MAPKinase, TNF, and NRF2 signaling pathways. Furthermore, the NRF2 signaling pathway which is upstream of the ATF3-MAPKinase signaling pathway, is also regulated by the mitochondrial oxidative stress. We are the first to report differential inflammatory characteristics of transcript variants 4 (ATF3-T4) and 5 (ATF3-T5) of the stress responsive gene ATF3 in HBMECs induced by postprandial TL. Specifically, our data indicates that ATF3-T4 predominantly regulates the TL-induced brain microvascular inflammation and TNF signaling. Both siRNAs of ATF3-T4 and ATF3-T5 suppress cells apoptosis and lipotoxic brain microvascular endothelial cells. These novel signaling pathways triggered by oxidative stress-responsive transcript variants, ATF3-T4 and ATF3-T5, in the brain microvascular inflammation induced by TGRL lipolysis products may contribute to pathophysiological processes of vascular dementia.

Rise of Deep Learning for Genomic, Proteomic, and Metabolomic Data Integration in Precision Medicine.

Machine learning (ML) is being ubiquitously incorporated into everyday products such as Internet search, email spam filters, product recommendations, image classification, and speech recognition. New approaches for highly integrated manufacturing and automation such as the Industry 4.0 and the Internet of things are also converging with ML methodologies. Many approaches incorporate complex artificial neural network architectures and are collectively referred to as deep learning (DL) applications. These methods have been shown capable of representing and learning predictable relationships in many diverse forms of data and hold promise for transforming the future of omics research and applications in precision medicine. Omics and electronic health record data pose considerable challenges for DL. This is due to many factors such as low signal to noise, analytical variance, and complex data integration requirements. However, DL models have already been shown capable of both improving the ease of data encoding and predictive model performance over alternative approaches. It may not be surprising that concepts encountered in DL share similarities with those observed in biological message relay systems such as gene, protein, and metabolite networks. This expert review examines the challenges and opportunities for DL at a systems and biological scale for a precision medicine readership.

Metabox: A Toolbox for Metabolomic Data Analysis, Interpretation and Integrative Exploration.

Similar to genomic and proteomic platforms, metabolomic data acquisition and analysis is becoming a routine approach for investigating biological systems. However, computational approaches for metabolomic data analysis and integration are still maturing. Metabox is a bioinformatics toolbox for deep phenotyping analytics that combines data processing, statistical analysis, functional analysis and integrative exploration of metabolomic data within proteomic and transcriptomic contexts. With the number of options provided in each analysis module, it also supports data analysis of other 'omic' families. The toolbox is an R-based web application, and it is freely available at http://kwanjeeraw.github.io/metabox/ under the GPL-3 license.

Metabolomics and Integrative Omics for the Development of Thai Traditional Medicine.

In recent years, interest in studies of traditional medicine in Asian and African countries has gradually increased due to its potential to complement modern medicine. In this review, we provide an overview of Thai traditional medicine (TTM) current development, and ongoing research activities of TTM related to metabolomics. This review will also focus on three important elements of systems biology analysis of TTM including analytical techniques, statistical approaches and bioinformatics tools for handling and analyzing untargeted metabolomics data. The main objective of this data analysis is to gain a comprehensive understanding of the system wide effects that TTM has on individuals. Furthermore, potential applications of metabolomics and systems medicine in TTM will also be discussed.