Large-scale Metabolomics Research Articles

Relationship of metabolites and metabolic ratios with schizophrenia: a mendelian randomization study

BackgroundThis study aims to investigate the causal relationship of human plasma metabolites and metabolic ratios with schizophrenia (SCZ).MethodsWe employed Mendelian Randomization (MR) approach to comprehensively analyze two large-scale metabolomics and schizophrenia Genome-Wide Association Study (GWAS) datasets, incorporating a total of 1091 metabolites and 309 metabolic ratios, with 52017 schizophrenia patients and 75889 healthy controls. The inverse variance-weighted (IVW) method was utilized to estimate the causal relationship between exposure and outcome. To provide a more comprehensive evaluation, additional Mendelian Randomization (MR) approaches were employed, including MR-Egger regression, weighted median, simple mode, and weighted mode methods. These analyses assessed the causal effects between blood metabolites, metabolic ratios, and schizophrenia. Tests for pleiotropy and heterogeneity were conducted. False Discovery Rate (FDR) correction was applied to account for multiple comparisons and heterogeneity, ensuring the robustness and reliability of our findings. Consistent with previous studies, an FDR threshold of < 0.2 was considered suggestive of a causal relationship, while an FDR of < 0.05 was considered to indicate a significant causal relationship.ResultsThe final results revealed that a significant causal association was found between the levels of two metabolites and schizophrenia, Alliin (OR = 0.915, 95%CI = 0.879–0.953, P = 1.93 × 10− 5, FDR = 0.013) was associated with a decreased risk of schizophrenia, N-actylcitrulline (OR = 1.058, 95%CI = 1.034–1.083, P = 1.4 × 10− 6, FDR = 0.002) was associated with increased risk of schizophrenia. When adjusting FDR to 0.2, the results showed that 4 metabolite levels and 2 metabolite ratios were suggestively causally associated with a reduced risk of schizophrenia including 2-aminooctanoate (OR = 0.904, 95%CI = 0.847–0.964, P = 0.002, FDR = 0.160), N-lactoylvaline (OR = 0.853, 95%CI = 0.775–0.938, P = 0.001,FDR = 0.122), X − 21310 (OR = 0.917, 95%CI = 0.866–0.971, P = 0.003,FDR = 0.195), X − 26111 (OR = 0.932, 95%CI = 0.890–0.976, P = 0.003,FDR = 0.189), Arachidonate (20:4n6) to oleate to vaccenate (18:1) ratio (OR = 0.945, 95%CI = 0.914–0.977, P = 8.2 × 10− 4, FDR = 0.104), and Citrulline to ornithine ratio (OR = 0.924, 95%CI = 0.881–0.969, P = 0.001, FDR = 0.122), while 4 metabolite levels and 2 metabolite ratios were suggestively causally associated with an increased risk of schizophrenia including N2, N5-diacetylornithine (OR = 1.090, 95%CI = 1.031–1.153, P = 0.003, FDR = 0.185), N − acetyl − 2−aminooctanoate (OR = 1.069, 95%CI=(1.027–1.114, P = 0.001, FDR = 0.127), N − acetyl − 2−aminoadipate (OR = 1.081, 95%CI = 1.030–1.133, P = 0.001, FDR = 0.128), X − 13844 (OR = 1.110, 95%CI = 1.036–1.190, P = 0.003, FDR = 0.196), X − 24556 (OR = 1.083, 95%CI = 1.036–1.132, P = 4.5 × 10− 4, FDR = 0.098), X − 24736 (OR = 1.065, 95%CI = 1.028–1.104, P = 5.6 × 10− 4, FDR = 0.098), N − acetylasparagine (OR = 1.048, 95%CI = 1.021–1.075, P = 4.5 × 10− 4, FDR = 0.098), N − acetylarginine (OR = 1.060, 95%CI = 1.028–1.092, P = 1.8 × 10− 4, FDR = 0.083), Cysteine to alanine ratio (OR = 1.086, 95%CI = 1.036–1.138, P = 6.5 × 10− 4, FDR = 0.101), and Benzoate to linoleoyl − arachidonoyl − glycerol (18:2 to 20:4) ratio (OR = 1.070, 95%CI = 1.025–1.117, P = 0.002, FDR = 0.158).ConclusionOur study results provide valuable insights for identifying diagnostic biomarkers related to schizophrenia and offer preliminary research findings for further exploration of the mechanisms linking schizophrenia and metabolism.

MetMiner: A user-friendly pipeline for large-scale plant metabolomics data analysis.

The utilization of metabolomics approaches to explore the metabolic mechanisms underlying plant fitness and adaptation to dynamic environments is growing, highlighting the need for an efficient and user-friendly toolkit tailored for analyzing the extensive datasets generated by metabolomics studies. Current protocols for metabolome data analysis often struggle with handling large-scale datasets or require programming skills. To address this, we present MetMiner (https://github.com/ShawnWx2019/MetMiner), a user-friendly, full-functionality pipeline specifically designed for plant metabolomics data analysis. Built on R shiny, MetMiner can be deployed on servers to utilize additional computational resources for processing large-scale datasets. MetMiner ensures transparency, traceability, and reproducibility throughout the analytical process. Its intuitive interface provides robust data interaction and graphical capabilities, enabling users without prior programming skills to engage deeply in data analysis. Additionally, we constructed and integrated a plant-specific mass spectrometry database into the MetMiner pipeline to optimize metabolite annotation. We have also developed MDAtoolkits, which include a complete set of tools for statistical analysis, metabolite classification, and enrichment analysis, to facilitate the mining of biological meaning from the datasets. Moreover, we propose an iterative weighted gene co-expression network analysis strategy for efficient biomarker metabolite screening in large-scale metabolomics data mining. In two case studies, we validated MetMiner's efficiency in data mining and robustness in metabolite annotation. Together, the MetMiner pipeline represents a promising solution for plant metabolomics analysis, providing a valuable tool for the scientific community to use with ease.

Validation of two LC[sbnd]HRMS methods for large-scale untargeted metabolomics of serum samples: Strategy to establish method fitness-for-purpose

Untargeted metabolomics by LCHRMS is a powerful tool to enhance our knowledge of pathophysiological processes. Whereas validation of a bioanalytical method is customary in most analytical chemistry fields, it is rarely performed for untargeted metabolomics. This study aimed to establish and validate an analytical platform for a long-term, clinical metabolomics study. Sample preparation was performed with an automated liquid handler and four analytical methods were developed and evaluated. The validation study spanned three batches with twelve runs using individual serum samples and various quality control samples. Data was acquired with untargeted acquisition and only metabolites identified at level 1 were evaluated. Validation parameters were set to evaluate key performance metrics relevant for the intended application: reproducibility, repeatability, stability, and identification selectivity, emphasizing dataset intrinsic variance. Concordance of semi-quantitative results between methods was evaluated to identify potential bias. Spearman rank correlation coefficients (rs) were calculated from individual serum samples. Of the four methods tested, two were selected for validation. A total of 47 and 55 metabolites (RPLC-ESI+- and HILIC-ESI−-HRMS, respectively) met specified validation criteria. Quality assurance involved system suitability testing, sample release, run release, and batch release. The median repeatability and within-run reproducibility as coefficient of variation% for metabolites that passed validation on RPLC-ESI+- and HILIC-ESI−-HRMS were 4.5 and 4.6, and 1.5 and 3.8, respectively. Metabolites that passed validation on RPLC-ESI+-HRMS had a median D-ratio of 1.91, and 89 % showed good signal intensity after ten-fold dilution. The corresponding numbers for metabolites with the HILIC-ESI−-HRMS method was 1.45 and 45 %, respectively. The rs median ({range}) for metabolites that passed validation on RPLC-ESI+- was 0.93 (N = 9 {0.69–0.98}) and on HILIC-ESI−-HRMS was 0.93 (N = 22 {0.55–1.00}). The validated methods proved fit-for-purpose and the laboratory thus demonstrated its capability to produce reliable results for a large-scale, untargeted metabolomics study. This validation not only bolsters the reliability of the assays but also significantly enhances the impact and credibility of the hypotheses generated from the studies. Therefore, this validation study serves as a benchmark in the documentation of untargeted metabolomics, potentially guiding future endeavors in the field.

Metabolic gene function discovery platform GeneMAP identifies SLC25A48 as necessary for mitochondrial choline import.

Organisms maintain metabolic homeostasis through the combined functions of small-molecule transporters and enzymes. While many metabolic components have been well established, a substantial number remains without identified physiological substrates. To bridge this gap, we have leveraged large-scale plasma metabolome genome-wide association studies (GWAS) to develop a multiomic Gene-Metabolite Association Prediction (GeneMAP) discovery platform. GeneMAP can generate accurate predictions and even pinpoint genes that are distant from the variants implicated by GWAS. In particular, our analysis identified solute carrier family 25 member 48 (SLC25A48) as a genetic determinant of plasma choline levels. Mechanistically, SLC25A48 loss strongly impairs mitochondrial choline import and synthesis of its downstream metabolite betaine. Integrative rare variant and polygenic score analyses in UK Biobank provide strong evidence that the SLC25A48 causal effects on human disease may in part be mediated by the effects of choline. Altogether, our study provides a discovery platform for metabolic gene function and proposes SLC25A48 as a mitochondrial choline transporter.

MetaboLink: A web application for Streamlined Processing and Analysis of Large-Scale Untargeted Metabolomics Data.

The post-processing and analysis of large-scale untargeted metabolomics data face significant challenges due to the intricate nature of correction, filtration, imputation, and normalization steps. Manual execution across various applications often leads to inefficiencies, human-induced errors, and inconsistencies within the workflow. Addressing these issues, we introduce MetaboLink, a novel web application designed to process LC-MS metabolomics datasets combining established methodologies and offering flexibility and ease of implementation. It offers visualization options for data interpretation, an interface for statistical testing, and integration with PolySTest for further tests and with VSClust for clustering analysis. Fully functional tool is publicly available at https://computproteomics.bmb.sdu.dk/Metabolomics/. The source code is available at https://github.com/anitamnd/MetaboLink and a detailed description of the app can be found at https://github.com/anitamnd/MetaboLink/wiki. A tutorial video can be found at https://youtu.be/ZM6j10S6Z8Q. Supplementary data are available at Bioinformatics online.

A large study of metabolomics reveals common and distinct metabolic biomarkers for type 2 diabetes, coronary heart disease, and stroke.

We aimed at examining the shared and unique associations of metabolites with multiple cardiometabolic diseases (CMD), i.e. type 2 diabetes (T2D), coronary heart disease (CHD) and stroke. In this study, a total of 168 plasma metabolites were measured by targeted high-throughput nuclear magnetic resonance spectroscopy among 98,162 participants free of T2D, CHD, and stroke at baseline. Cox proportional hazard models estimated hazard ratios for one SD increase in metabolite concentration levels, and false discovery rate (at 10%) was used to correct for multiple comparisons. Over 12.1 years of follow-up on average, 3,463 T2D, 6,186 CHD, and 1,892 stroke events were recorded. Most lipoprotein metabolites were associated with risks of T2D and CHD but not with the risk of stroke, with stronger associations for T2D than for CHD. Phospholipids within intermediate-density lipoprotein or large low-density lipoprotein particles showed positive associations with CHD and inverse associations with T2D. Metabolites indicating very small very low-density lipoprotein, histidine, creatinine, albumin, and glycoprotein acetyls were associated with risks of all three conditions. This large-scale metabolomics study revealed common and distinct metabolic biomarkers for T2D, CHD and stroke, providing instrumental information to possibly implement precision medicine for preventing and treating these conditions.

Common data models to streamline metabolomics processing and annotation, and implementation in a Python pipeline.

To standardize metabolomics data analysis and facilitate future computational developments, it is essential to have a set of well-defined templates for common data structures. Here we describe a collection of data structures involved in metabolomics data processing and illustrate how they are utilized in a full-featured Python-centric pipeline. We demonstrate the performance of the pipeline, and the details in annotation and quality control using large-scale LC-MS metabolomics and lipidomics data and LC-MS/MS data. Multiple previously published datasets are also reanalyzed to showcase its utility in biological data analysis. This pipeline allows users to streamline data processing, quality control, annotation, and standardization in an efficient and transparent manner. This work fills a major gap in the Python ecosystem for computational metabolomics.

Consistency of metabolite associations with measured glomerular filtration rate in children and adults.

There is interest in identifying novel filtration markers that lead to more accurate GFR estimates than current markers (creatinine and cystatin C) and are more consistent across demographic groups. We hypothesize that large-scale metabolomics can identify serum metabolites that are strongly influenced by glomerular filtration rate (GFR) and are more consistent across demographic variables than creatinine, which would be promising filtration markers for future investigation. We evaluated the consistency of associations between measured GFR (mGFR) and 887 common, known metabolites quantified by an untargeted chromatography- and spectroscopy-based metabolomics platform (Metabolon) performed on frozen blood samples from 580 participants in Chronic Kidney Disease in Children (CKiD), 674 participants in Modification of Diet in Renal Disease (MDRD) Study and 962 participants in African American Study of Kidney Disease and Hypertension (AASK). We evaluated metabolite-mGFR correlation association with metabolite class, molecular weight, assay platform and measurement coefficient of variation (CV). Among metabolites with strong negative correlations with mGFR (r < -0.5), we assessed additional variation by age (height in children), sex, race and body mass index (BMI). A total of 561 metabolites (63%) were negatively correlated with mGFR. Correlations with mGFR were highly consistent across study, sex, race and BMI categories (correlation of metabolite-mGFR correlations between 0.88 and 0.95). Amino acids, carbohydrates and nucleotides were more often negatively correlated with mGFR compared with lipids, but there was no association with metabolite molecular weight, liquid chromatography/mass spectrometry platform and measurement CV. Among 114 metabolites with strong negative associations with mGFR (r < -0.5), 27 were consistently not associated with age (height in children), sex or race. The majority of metabolite-mGFR correlations were negative and consistent across sex, race, BMI and study. Metabolites with consistent strong negative correlations with mGFR and non-association with demographic variables may represent candidate markers to improve estimation of GFR.

Postmortem metabolomics as a high-throughput cause-of-death screening tool for human death investigations

Autopsy rates are declining globally, impacting cause-of-death (CoD) diagnoses and quality control. Postmortem metabolomics was evaluated for CoD screening using 4,282 human cases, encompassing CoD groups: acidosis, drug intoxication, hanging, ischemic heart disease (IHD), and pneumonia. Cases were split 3:1 into training and test sets. High-resolution mass spectrometry data from femoral blood were analyzed via orthogonal-partial least squares discriminant analysis (OPLS-DA) to discriminate CoD groups. OPLS-DA achieved an R2= 0.52 and Q2= 0.30, with true-positive prediction rates of 68% and 65% for training and test sets, respectively, across all groups. Specificity-optimized thresholds predicted 56% of test cases with a unique CoD, average 45% sensitivity, and average 96% specificity. Prediction accuracies varied: 98.7% for acidosis, 80.5% for drug intoxication, 81.6% for hanging, 73.1% for IHD, and 93.6% for pneumonia. This study demonstrates the potential of large-scale postmortem metabolomics for CoD screening, offering high specificity and enhancing throughput and decision-making in human death investigations.

The underappreciated diversity of bile acid modifications

The repertoire of modifications to bile acids and related steroidal lipids by host and microbial metabolism remains incompletely characterized. To address this knowledge gap, we created a reusable resource of tandem mass spectrometry (MS/MS) spectra by filtering 1.2 billion publicly available MS/MS spectra for bile-acid-selective ion patterns. Thousands of modifications are distributed throughout animal and human bodies as well as microbial cultures. We employed this MS/MS library to identify polyamine bile amidates, prevalent in carnivores. They are present in humans, and their levels alter with a diet change from a Mediterranean to a typical American diet. This work highlights the existence of many more bile acid modifications than previously recognized and the value of leveraging public large-scale untargeted metabolomics data to discover metabolites. The availability of a modification-centric bile acid MS/MS library will inform future studies investigating bile acid roles in health and disease.

Reducing Quantitative Uncertainty Caused by Data Processing in Untargeted Metabolomics.

Processing liquid chromatography-mass spectrometry-based metabolomics data using computational programs often introduces additional quantitative uncertainty, termed computational variation in a previous work. This work develops a computational solution to automatically recognize metabolic features with computational variation in a metabolomics data set. This tool, AVIR (short for "Accurate eValuation of alIgnment and integRation"), is a support vector machine-based machine learning strategy (https://github.com/HuanLab/AVIR). The rationale is that metabolic features with computational variation have a poor correlation between chromatographic peak area and peak height-based quantifications across the samples in a study. AVIR was trained on a set of 696 manually curated metabolic features and achieved an accuracy of 94% in a 10-fold cross-validation. When tested on various external data sets from public metabolomics repositories, AVIR demonstrated an accuracy range of 84%-97%. Finally, tested on a large-scale metabolomics study, AVIR clearly indicated features with computational variation and thus guided us to manually correct them. Our results show that 75.3% of the samples with computational variation had a relative intensity difference of over 20% after correction. This demonstrates the critical role of AVIR in reducing computational variation to improve quantitative certainty in untargeted metabolomics analysis.

Ion entropy and accurate entropy-based FDR estimation in metabolomics.

Accurate metabolite annotation and false discovery rate (FDR) control remain challenging in large-scale metabolomics. Recent progress leveraging proteomics experiences and interdisciplinary inspirations has provided valuable insights. While target-decoy strategies have been introduced, generating reliable decoy libraries is difficult due to metabolite complexity. Moreover, continuous bioinformatics innovation is imperative to improve the utilization of expanding spectral resources while reducing false annotations. Here, we introduce the concept of ion entropy for metabolomics and propose two entropy-based decoy generation approaches. Assessment of public databases validates ion entropy as an effective metric to quantify ion information in massive metabolomics datasets. Our entropy-based decoy strategies outperform current representative methods in metabolomics and achieve superior FDR estimation accuracy. Analysis of 46 public datasets provides instructive recommendations for practical application.

Enhancing Metabolome Annotation by Electron Impact Excitation of Ions from Organics-Molecular Networking.

Liquid chromatography-high-resolution mass spectrometry (LC-HRMS) is widely used in untargeted metabolomics, but large-scale and high-accuracy metabolite annotation remains a challenge due to the complex nature of biological samples. Recently introduced electron impact excitation of ions from organics (EIEIO) fragmentation can generate information-rich fragment ions. However, effective utilization of EIEIO tandem mass spectrometry (MS/MS) is hindered by the lack of reference spectral databases. Molecular networking (MN) shows great promise in large-scale metabolome annotation, but enhancing the correlation between spectral and structural similarity is essential to fully exploring the benefits of MN annotation. In this study, a novel approach was proposed to enhance metabolite annotation in untargeted metabolomics using EIEIO and MN. MS/MS spectra were acquired in EIEIO and collision-induced dissociation (CID) modes for over 400 reference metabolites. The study revealed a stronger correlation between the EIEIO spectra and metabolite structure. Moreover, the EIEIO spectral network outperformed the CID spectral network in capturing structural analogues. The annotation performance of the structural similarity network for untargeted LC-MS/MS was evaluated. For the spiked NIST SRM 1950 human plasma, the annotation coverage and accuracy were 72.94 and 74.19%, respectively. A total of 2337 metabolite features were successfully annotated in NIST SRM 1950 human plasma, which was twice that of LC-CID MS/MS. Finally, the developed method was applied to investigate prostate cancer. A total of 87 significantly differential metabolites were annotated. This study combining EIEIO and MN makes a valuable contribution to improving metabolome annotation.

Renal Mitochondrial ATP Transporter Ablation Ameliorates Obesity-Induced CKD.

This study sheds light on the central role of adenine nucleotide translocase 2 (ANT2) in the pathogenesis of obesity-induced CKD. Our data demonstrate that ANT2 depletion in renal proximal tubule cells (RPTCs) leads to a shift in their primary metabolic program from fatty acid oxidation to aerobic glycolysis, resulting in mitochondrial protection, cellular survival, and preservation of renal function. These findings provide new insights into the underlying mechanisms of obesity-induced CKD and have the potential to be translated toward the development of targeted therapeutic strategies for this debilitating condition. The impairment in ATP production and transport in RPTCs has been linked to the pathogenesis of obesity-induced CKD. This condition is characterized by kidney dysfunction, inflammation, lipotoxicity, and fibrosis. In this study, we investigated the role of ANT2, which serves as the primary regulator of cellular ATP content in RPTCs, in the development of obesity-induced CKD. We generated RPTC-specific ANT2 knockout ( RPTC-ANT2-/- ) mice, which were then subjected to a 24-week high-fat diet-feeding regimen. We conducted comprehensive assessment of renal morphology, function, and metabolic alterations of these mice. In addition, we used large-scale transcriptomics, proteomics, and metabolomics analyses to gain insights into the role of ANT2 in regulating mitochondrial function, RPTC physiology, and overall renal health. Our findings revealed that obese RPTC-ANT2-/- mice displayed preserved renal morphology and function, along with a notable absence of kidney lipotoxicity and fibrosis. The depletion of Ant2 in RPTCs led to a fundamental rewiring of their primary metabolic program. Specifically, these cells shifted from oxidizing fatty acids as their primary energy source to favoring aerobic glycolysis, a phenomenon mediated by the testis-selective Ant4. We propose a significant role for RPTC-Ant2 in the development of obesity-induced CKD. The nullification of RPTC-Ant2 triggers a cascade of cellular mechanisms, including mitochondrial protection, enhanced RPTC survival, and ultimately the preservation of kidney function. These findings shed new light on the complex metabolic pathways contributing to CKD development and suggest potential therapeutic targets for this condition.

How to Better Explain and Develop Evolution Today

Studies over the past two decades have provided important information on evolution. An essential discipline gradually formed called evolutionary developmental biology, or evo-devo combines the principles of evolutionary biology and developmental biology to investigate the genetic and environmental factors that affect the development of organisms and how these elements affect evolutionary change. In this discipline, evolution can be explained by genetics, cell biology, and developmental biology. These fields provide crucial insights into how genetic variation, cellular processes, and developmental mechanisms contribute to population changes over time. With the popularity of emerging information technologies such as mobile Internet and big data, the global digital economy is booming. The development of gene sequencing technology has led to the rapid growth of genomic data, and the combination of large-scale genomics, metabolomics, and functional research systems, known as “big data” and bioinformatics, has brought about a “technological revolution” in evolutionary research. Therefore, enhancing the interconnection between scientific research in various fields is extremely necessary and will be effective.

A large-scale, targeted metabolomics method for the analysis and quantification of metabolites in human plasma via liquid chromatography-mass spectrometry

Metabolomics is the study of small molecules, primarily metabolites, that are produced during metabolic processes. Analysis of the composition of an organism's metabolome can yield useful information about an individual's health status at any given time. In recent years, the development of large-scale, targeted metabolomic methods has allowed for the analysis of biological samples using analytical techniques such as LC-MS/MS. This paper presents a large-scale metabolomics method for analysis of biological samples, with a focus on quantification of metabolites found in blood plasma. The method comprises a 10-min chromatographic separation using HILIC and RP stationary phases combined with positive and negative electrospray ionization in order to maximize metabolome coverage. Complete analysis of a single sample can be achieved in as little as 40 min using the two columns and dual modes of ionization. With 540 metabolites and the inclusion of over 200 analytical standards, this method is comprehensive and quantitatively robust when compared to current targeted metabolomics methods. This study uses a large-scale evaluation of metabolite recovery from plasma that enables absolute quantification of metabolites by correcting for analyte loss throughout processes such as extraction, handling, or storage. In addition, the method was applied to plasma collected from adjuvant breast cancer patients to confirm the suitability of the method to clinical samples.

Novel Peak Shift Correction Method Based on the Retention Index for Peak Alignment in Untargeted Metabolomics.

Peak alignment is a crucial step in liquid chromatography-mass spectrometry (LC-MS)-based large-scale untargeted metabolomics workflows, as it enables the integration of metabolite peaks across multiple samples, which is essential for accurate data interpretation. Slight differences or fluctuations in chromatographic separation conditions, however, can cause the chromatographic retention time (RT) shift between consecutive analyses, ultimately affecting the accuracy of peak alignment between samples. Here, we introduce a novel RT shift correction method based on the retention index (RI) and apply it to peak alignment. We synthesized a series of N-acyl glycine (C2-C23) homologues via the amidation reaction between glycine with normal saturated fatty acids (C2-C23) as calibrants able to respond proficiently in both mass spectrometric positive- and negative-ion modes. Using these calibrants, we established an N-acyl glycine RI system. This RI system is capable of covering a broad chromatographic space and addressing chromatographic RT shift caused by variations in flow rate, gradient elution, instrument systems, and LC separation columns. Moreover, based on the RI system, we developed a peak shift correction model to enhance peak alignment accuracy. Applying the model resulted in a significant improvement in the accuracy of peak alignment from 15.5 to 80.9% across long-term data spanning a period of 157 days. To facilitate practical application, we developed a Python-based program, which is freely available at https://github.com/WHU-Fenglab/RI-based-CPSC.

Denoising Autoencoder Normalization for Large-Scale Untargeted Metabolomics by Gas Chromatography-Mass Spectrometry.

Large-scale metabolomics assays are widely used in epidemiology for biomarker discovery and risk assessments. However, systematic errors introduced by instrumental signal drifting pose a big challenge in large-scale assays, especially for derivatization-based gas chromatography-mass spectrometry (GC-MS). Here, we compare the results of different normalization methods for a study with more than 4000 human plasma samples involved in a type 2 diabetes cohort study, in addition to 413 pooled quality control (QC) samples, 413 commercial pooled plasma samples, and a set of 25 stable isotope-labeled internal standards used for every sample. Data acquisition was conducted across 1.2 years, including seven column changes. In total, 413 pooled QC (training) and 413 BioIVT samples (validation) were used for normalization comparisons. Surprisingly, neither internal standards nor sum-based normalizations yielded median precision of less than 30% across all 563 metabolite annotations. While the machine-learning-based SERRF algorithm gave 19% median precision based on the pooled quality control samples, external cross-validation with BioIVT plasma pools yielded a median 34% relative standard deviation (RSD). We developed a new method: systematic error reduction by denoising autoencoder (SERDA). SERDA lowered the median standard deviations of the training QC samples down to 16% RSD, yielding an overall error of 19% RSD when applied to the independent BioIVT validation QC samples. This is the largest study on GC-MS metabolomics ever reported, demonstrating that technical errors can be normalized and handled effectively for this assay. SERDA was further validated on two additional large-scale GC-MS-based human plasma metabolomics studies, confirming the superior performance of SERDA over SERRF or sum normalizations.

How metabolism and development are intertwined in space and time.

Developmental transitions, occurring throughout the life cycle of plants, require precise regulation of metabolic processes to generate the energy and resources necessary for the committed growth processes. In parallel, the establishment of new cells, tissues, and even organs, alongside their differentiation provoke profound changes in metabolism. It is increasingly being recognized that there is a certain degree of feedback regulation between the components and products of metabolic pathways and developmental regulators. The generation of large-scale metabolomics datasets during developmental transitions, in combination with molecular genetic approaches has helped to further our knowledge on the functional importance of metabolic regulation of development. In this perspective article, we provide insights into studies that elucidate interactions between metabolism and development at the temporal and spatial scales. We additionally discuss how this influences cell growth-related processes. We also highlight how metabolic intermediates function as signaling molecules to direct plant development in response to changing internal and external conditions.

COVRECON: automated integration of genome- and metabolome-scale network reconstruction and data-driven inverse modeling of metabolic interaction networks.

One central goal of systems biology is to infer biochemical regulations from large-scale OMICS data. Many aspects of cellular physiology and organismal phenotypes can be understood as results of metabolic interaction network dynamics. Previously, we have proposed a convenient mathematical method, which addresses this problem using metabolomics data for the inverse calculation of biochemical Jacobian matrices revealing regulatory checkpoints of biochemical regulations. The proposed algorithms for this inference are limited by two issues: they rely on structural network information that needs to be assembled manually, and they are numerically unstable due to ill-conditioned regression problems for large-scale metabolic networks. To address these problems, we developed a novel regression loss-based inverse Jacobian algorithm, combining metabolomics COVariance and genome-scale metabolic RECONstruction, which allows for a fully automated, algorithmic implementation of the COVRECON workflow. It consists of two parts: (i) Sim-Network and (ii) inverse differential Jacobian evaluation. Sim-Network automatically generates an organism-specific enzyme and reaction dataset from Bigg and KEGG databases, which is then used to reconstruct the Jacobian's structure for a specific metabolomics dataset. Instead of directly solving a regression problem as in the previous workflow, the new inverse differential Jacobian is based on a substantially more robust approach and rates the biochemical interactions according to their relevance from large-scale metabolomics data. The approach is illustrated by in silico stochastic analysis with differently sized metabolic networks from the BioModels database and applied to a real-world example. The characteristics of the COVRECON implementation are that (i) it automatically reconstructs a data-driven superpathway model; (ii) more general network structures can be investigated, and (iii) the new inverse algorithm improves stability, decreases computation time, and extends to large-scale models. The code is available in the website https://bitbucket.org/mosys-univie/covrecon.