Random forest-based imputation outperforms other methods for imputing LC-MS metabolomics data: a comparative study

Integrated metabolic-transcriptomic network identifies immunometabolic modulations in human macrophages

Macrophages exhibit diverse phenotypes and respond flexibly to environmental cues through metabolic remodeling. In this study, we present a comprehensive multi-omics dataset integrating intra- and extracellular metabolomes with transcriptomic data to investigate the metabolic impact on human macrophage function. Our analysis establishes a metabolite-gene correlation network that characterizes macrophage activation. We find that the concurrent inhibition of tryptophan catabolism by IDO1 and IL4I1 inhibitors suppresses the macrophage pro-inflammatory response, whereas single inhibition leads to pro-inflammatory activation. We find that a subset of anti-inflammatory macrophages activated by Fc receptor signaling promotes glycolysis, challenging the conventional concept of reduced glycolysis preference in anti-inflammatory macrophages. We demonstrate that cholesterol accumulation suppresses macrophage IFN-γ responses. Our integrated network enables the discovery of immunometabolic features, provides insights into macrophage functional metabolic reprogramming, and offers valuable resources for researchers exploring macrophage immunometabolic characteristics and potential therapeutic targets for immune-related disorders.

The HuMet Repository: Watching human metabolism at work

Metabolism oscillates between catabolic and anabolic states depending on food intake, exercise, or stresses that change a multitude of metabolic pathways simultaneously. We present the HuMet Repository for exploring dynamic metabolic responses to oral glucose/lipid loads, mixed meals, 36-h fasting, exercise, and cold stress in healthy subjects. Metabolomics data from blood, urine, and breath of 15 young, healthy men at up to 56 time points are integrated and embedded within an interactive web application, enabling researchers with and without computational expertise to search, visualize, analyze, and contextualize the dynamic metabolite profiles of 2,656 metabolites acquired on multiple platforms. With examples, we demonstrate the utility of the resource for research into the dynamics of human metabolism, highlighting differences and similarities in systemic metabolic responses across challenges and the complementarity of metabolomics platforms. The repository, providing a reference for healthy metabolite changes to six standardized physiological challenges, is freely accessible through a web portal.

Genetic Architecture and Analysis Practices of Circulating Metabolites in the NHLBI Trans-Omics for Precision Medicine (TOPMed) Program

Circulating metabolite levels partly reflect the state of human health and diseases, and can be impacted by genetic determinants. Hundreds of loci associated with circulating metabolites have been identified; however, most findings focus on predominantly European ancestry or single study analyses. Leveraging the rich metabolomics resources generated by the NHLBI Trans-Omics for Precision Medicine (TOPMed) Program, we harmonized and accessibly cataloged 1,729 circulating metabolites among 25,058 ancestrally-diverse samples. We provided recommendations for outlier and imputation handling to process metabolite data, as well as a general analytical framework. We further performed a pooled analysis following our practical recommendations and discovered 1,778 independent loci associated with 667 metabolites. Among 108 novel locus - metabolite pairs, we detected not only novel loci within previously implicated metabolite associated genes, but also novel genes (such as GAB3 and VSIG4 located in the X chromosome) that have putative roles in metabolic regulation. In the sex-stratified analysis, we revealed 85 independent locus-metabolite pairs with evidence of sexual dimorphism, including well-known metabolic genes such as FADS2 , D2HGDH , SUGP1 , UTG2B17 , strongly supporting the importance of exploring sex difference in the human metabolome. Taken together, our study depicted the genetic contribution to circulating metabolite levels, providing additional insight into the understanding of human health.

Racial and ethnic disparities in post-liver transplant outcomes for patients with acute-on-chronic liver failure: An analysis of the UNOS database

BACKGROUND

The incidence of hospitalisations related to acute-on-chronic liver failure (ACLF) is increasing. Liver transplantation (LT) remains the definitive treatment for the condition.

AIM

To evaluate the influence of race and ethnicity on LT outcomes in ACLF.

METHODS

We conducted a retrospective analysis utilising LT data from the United Network for Organ Sharing (UNOS) database. White patients served as the control group and patients of other races were compared at each ACLF grade. The primary outcomes assessed were graft failure and all-cause mortality.

RESULTS

Blacks exhibited a higher all-cause mortality (Grade 1: aHR 1.36, 95% CI 1.18-1.57, p < 0.001; Grade 2: aHR 1.27, 95% CI 1.08-1.48, p = 0.003; Grade 3: aHR 1.19, 95% CI 1.04-1.37, p = 0.01) and graft failure (Grade 1: aHR 2.05, 95% CI 1.58-2.67, p < 0.001; Grade 2: aHR 1.91, 95% CI 1.43-2.54, p < 0.001; Grade 3: aHR 1.50, 95% CI 1.15-1.96, p = 0.002). Hispanics experienced a lower all-cause mortality at grades 1 and 3 (Grade 1: aHR 0.83, 95% CI 0.72-0.96, p = 0.01; Grade 3: aHR 0.80, 95% CI 0.70-0.91, p < 0.001) and Asians with severe ACLF demonstrated decreased all-cause mortality (Grade 3: aHR 0.55, 95% CI 0.42-0.73, p < 0.001).

CONCLUSION

Black patients experienced the poorest outcomes and Hispanic and Asian patients demonstrated more favourable outcomes compared to Whites.

Comprehensive Overview of Bottom-Up Proteomics Using Mass Spectrometry

Proteomics is the large scale study of protein structure and function from biological systems through protein identification and quantification. "Shotgun proteomics" or "bottom-up proteomics" is the prevailing strategy, in which proteins are hydrolyzed into peptides that are analyzed by mass spectrometry. Proteomics studies can be applied to diverse studies ranging from simple protein identification to studies of proteoforms, protein-protein interactions, protein structural alterations, absolute and relative protein quantification, post-translational modifications, and protein stability. To enable this range of different experiments, there are diverse strategies for proteome analysis. The nuances of how proteomic workflows differ may be challenging to understand for new practitioners. Here, we provide a comprehensive overview of different proteomics methods. We cover from biochemistry basics and protein extraction to biological interpretation and orthogonal validation. We expect this Review will serve as a handbook for researchers who are new to the field of bottom-up proteomics.

Predicting progression-free survival in patients with epithelial ovarian cancer using an interpretable random forest model

Prognostic models play a crucial role in providing personalised risk assessment, guiding treatment decisions, and facilitating the counselling of patients with cancer. However, previous imaging-based artificial intelligence models of epithelial ovarian cancer lacked interpretability. In this study, we aimed to develop an interpretable machine-learning model to predict progression-free survival in patients with epithelial ovarian cancer using clinical variables and radiomics features. A total of 102 patients with epithelial ovarian cancer who underwent contrast-enhanced computed tomography scans were enrolled in this retrospective study. Pre-surgery clinical data, including age, performance status, body mass index, tumour stage, venous blood cancer antigen-125 (CA125) level, white blood cell count, neutrophil count, red blood cell count, haemoglobin level, and platelet count, were obtained from medical records. The volume of interest for each tumour was manually delineated slice-by-slice along the boundary. A total of 2074 radiomic features were extracted from the pre- and post-contrast computed tomography images. Optimal radiomic features were selected using the Least Absolute Shrinkage and Selection Operator logistic regression. Multivariate Cox analysis was performed to identify independent predictors of three-year progression-free survival. The random forest algorithm developed radiomic and combined models using four-fold cross-validation. Finally, the Shapley additive explanation algorithm was applied to interpret the predictions of the combined model. Multivariate Cox analysis identified CA-125 levels (P = 0.015), tumour stage (P = 0.019), and Radscore (P < 0.001) as independent predictors of progression-free survival. The combined model based on these factors achieved an area under the curve of 0.812 (95 % confidence interval: 0.802-0.822) in the training cohort and 0.772 (95 % confidence interval: 0.727-0.817) in the validation cohort. The most impactful features on the model output were Radscore, followed by tumour stage and CA-125. In conclusion, the Shapley additive explanation-based interpretation of the prognostic model enables clinicians to understand the reasoning behind predictions better.

Plasma and milk metabolomics profiles in dairy cows with subclinical and clinical ketosis

Ketosis, a commonly observed energy metabolism disorder in dairy cows during the peripartal period, is distinguished by increased concentrations of BHB in the blood. This condition has a negative impact on milk production and quality, causing financial losses. An untargeted metabolomics approach was performed on plasma samples from cows between 5 and 7 DIM diagnosed as controls (CON; BHB <1.2 mM, n = 30), subclinically ketotic (SCK; 1.2 < BHB <3.0 mM, n = 30), or clinically ketotic (CK; BHB >3.0 mM, n = 30). Cows were selected from a commercial farm of 214 Holstein cows (average 305-d yield in the previous lactation of 35.42 ± 7.23 kg/d; parity, 2.41 ± 1.12; BCS, 3.1 ± 0.45). All plasma and milk samples (n = 90) were subjected to liquid chromatography-MS-based metabolomic analysis. Statistical analyses were performed using GraphPad Prism 8.0, MetaboAnalyst 4.0, and R version 4.1.3. Compared with the CON group, both SCK and CK groups had greater milk fat, freezing point, and fat-to-protein ratio, as well as lower milk protein, lactose, solids-not-fat, and milk density. Within 21 d after calving, compared with CON, the SCK group experienced a reduction of 2.65 kg/d in milk yield, while the CK group experienced a decrease of 7.7 kg/d. Untargeted metabolomics analysis facilitated the annotation of a total of 5,259 and 8,423 metabolites in plasma and milk. Differentially affected metabolites were screened in CON versus SCK, CON versus CK, and SCK versus CK (unpaired t-test, false discovery rate <0.05; and absolute value of log(2)-fold change >1.5). A total of 1,544 and 1,888 differentially affected metabolites were detected in plasma and milk. In plasma, glycerophospholipid metabolism, pyrimidine metabolism, tryptophan metabolism, sphingolipid metabolism, amino sugar and nucleotide sugar metabolism, phenylalanine metabolism, and steroid hormone biosynthesis were identified as important pathways. Weighted gene co-expression network analysis (WGCNA) indicated that tryptophan metabolism is a key pathway associated with the occurrence and development of ketosis. Increases in 5-hydroxytryptophan and decreases in kynurenine and 3-indoleacetic acid in SCK and CK were suggestive of an impact at the gut level. The decrease of most glycerophospholipids indicated that ketosis is associated with disordered lipid metabolism. For milk, pyrimidine metabolism, purine metabolism, pantothenate and CoA biosynthesis, amino sugar and nucleotide sugar metabolism, nicotinate and nicotinamide metabolism, sphingolipid metabolism, and fatty acid degradation were identified as important pathways. The WGCNA indicated that purine and pyrimidine metabolism in plasma was highly correlated with milk yield during the peripartal period. Alterations in purine and pyrimidine metabolism characterized ketosis, with lower levels of these metabolites in both milk and blood underscoring reduced efficiency in nitrogen metabolism. Our results may help to establish a foundation for future research investigating mechanisms responsible for the occurrence and development of ketosis in peripartal cows.

Untargeted metabolomics and machine learning unveil quality and authenticity interactions in grated Parmigiano Reggiano PDO cheese

The chemical composition of Parmigiano Reggiano (PR) hard cheese can be significantly affected by different factors across the dairy supply chain, including ripening, altimetric zone, and rind inclusion levels in grated hard cheeses. The present study proposes an untargeted metabolomics approach combined with machine learning chemometrics to evaluate the combined effect of these three critical parameters. Specifically, ripening was found to exert a pivotal role in defining the signature of PR cheeses, with amino acids and lipid derivatives that exhibited their role as key discriminant compounds. In parallel, a random forest classifier was used to predict the rind inclusion levels (> 18%) in grated cheeses and to authenticate the specific effect of altimetry dairy production, achieving a high prediction ability in both model performances (i.e., ∼60% and > 90%, respectively). Overall, these results open a novel perspective to identifying quality and authenticity markers metabolites in cheese.

Metabolomics signatures of sweetened beverages and added sugar are related to anthropometric measures of adiposity in young individuals: results from a cohort study

BACKGROUND

The associations of sweetened beverages (SBs) and added sugar (AS) intake with adiposity are still debated. Metabolomics could provide insights into the mechanisms linking their intake to adiposity.

OBJECTIVES

We aimed to identify metabolomics biomarkers of intake of low- and no-calorie sweetened beverages (LNCSBs), sugar-sweetened beverages (SSBs), and ASs and to investigate their associations with body mass index, body fat percentage, and waist circumference.

METHODS

We analyzed 3 data sets from the Dortmund Nutritional and Anthropometric Longitudinally Designed (DONALD) cohort study, of children who provided 2 urine samples (n = 297), adolescents who provided a single urine sample (n = 339), and young adults who provided a single plasma sample (n = 195). Urine and plasma were analyzed using untargeted metabolomics. Dietary intakes were assessed using 3-d weighed dietary records. The random forest, partial least squares, and least absolute shrinkage and selection operator were jointly used for metabolite selection. We examined associations of intakes with metabolites and anthropometric measures using linear and mixed-effects regression.

RESULTS

In adolescents, LNCSB were positively associated with acesulfame (β: 0.0012; 95% confidence interval [CI]: 0.0006, 0.0019) and saccharin (β: 0.0009; 95% CI: 0.0002, 0.0015). In children, the association was observed with saccharin (β: 0.0016; 95% CI: 0.0005, 0.0027). In urine and plasma, SSBs were positively associated with 1-methylxanthine (β: 0.0005; 95% CI: 0.0003, 0.0008; and β: 0.0010, 95% CI 0.0004, 0.0015, respectively) and 5-acetylamino-6-amino-3-methyluracil (β: 0.0005; 95% CI: 0.0002, 0.0008; and β: 0.0009; 95% CI: 0.0003, 0.0014, respectively). AS was associated with urinary sucrose (β: 0.0095; 95% CI: 0.0069, 0.0121) in adolescents. Some of the food-related metabolomics profiles were also associated with adiposity measures.

CONCLUSIONS

We identified SBs- and AS-related metabolites, which may be important for understanding the interplay between these intakes and adiposity in young individuals.

A principal component analysis of metabolome and cognitive decline among Japanese older adults: cross-sectional analysis using Tohoku Medical Megabank Cohort Study

BackgroundDementia is the leading cause of disability and imposes a significant burden on society. Previous studies have suggested an association between metabolites and cognitive decline. Although the metabolite composition differs between Western and Asian populations, studies targeting Asian populations remain scarce.MethodsThis cross-sectional study used data from a cohort survey of community-dwelling older adults aged ≥ 60 years living in Miyagi, Japan, conducted by Tohoku Medical Megabank Organization between 2013 and 2016. Forty-three metabolite variables quantified using nuclear magnetic resonance spectroscopy were used as explanatory variables. Dependent variable was the presence of cognitive decline (≤ 23 points), assessed by the Mini-Mental State Examination. Principal component (PC) analysis was performed to reduce the dimensionality of metabolite variables, followed by logistic regression analysis to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for cognitive decline.ResultsA total of 2,940 participants were included (men: 49.0%, mean age: 67.6 years). Among them, 1.9% showed cognitive decline. The first 12 PC components (PC1-PC12) accounted for 71.7% of the total variance. Multivariate analysis showed that PC1, which mainly represented essential amino acids, was associated with lower odds of cognitive decline (OR = 0.89; 95% CI, 0.80-0.98). PC2, which mainly included ketone bodies, was associated with cognitive decline (OR = 1.29; 95% CI, 1.11-1.51). PC3, which included amino acids, was associated with lower odds of cognitive decline (OR = 0.81; 95% CI, 0.66-0.99).ConclusionAmino acids are protectively associated with cognitive decline, whereas ketone metabolites are associated with higher odds of cognitive decline.

Inpatient Cost Burdens of Treating Chronic Hepatitis B in US Hospitals: A Weighted Analysis of a National Database

BACKGROUND AND AIMS

This study evaluates the cost burdens of inpatient care for chronic hepatitis B (CHB). We aimed to stratify the patients based on the presence of cirrhosis and conduct subgroup analyses on patient demographics and medical characteristics.

METHODS

The 2016-2019 National Inpatient Sample was used to select individuals diagnosed with CHB. The weighted charge estimates were derived and converted to admission costs, adjusting for inflation to the year 2016, and presented in United States Dollars. These adjusted values were stratified using select patient variables. To assess the goodness-of-fit for each trend, we graphed the data across the respective years, expressed in a chronological sequence with format (R2, p-value). Analysis of CHB patients was carried out in three groups: the composite CHB population, the subset of patients with cirrhosis, and the subset of patients without cirrhosis.

RESULTS

From 2016 to 2019, the total costs of hospitalizations in CHB patients were $603.82, $737.92, $758.29, and $809.01 million dollars from 2016 to 2019, respectively. We did not observe significant cost trends in the composite CHB population or in the cirrhosis and non-cirrhosis cohorts. However, we did find rising costs associated with age older than 65 (0.97, 0.02), white race (0.98, 0.01), Hispanic ethnicity (1.00, 0.001), and Medicare coverage (0.95, 0.02), the significance of which persisted regardless of the presence of cirrhosis. Additionally, inpatients without cirrhosis who had comorbid metabolic dysfunction-associated steatotic liver disease (MASLD) were also observed to have rising costs (0.96, 0.02).

CONCLUSIONS

We did not find a significant increase in overall costs with CHB inpatients, regardless of the presence of cirrhosis. However, certain groups are more susceptible to escalating costs. Therefore, increased screening and nuanced vaccination planning must be optimized in order to prevent and mitigate these growing cost burdens on vulnerable populations.

omicsMIC: a comprehensive benchmarking platform for robust comparison of imputation methods in mass spectrometry-based omics data

Mass spectrometry is a powerful and widely used tool for generating proteomics, lipidomics and metabolomics profiles, which is pivotal for elucidating biological processes and identifying biomarkers. However, missing values in mass spectrometry-based omics data may pose a critical challenge for the comprehensive identification of biomarkers and elucidation of the biological processes underlying human complex disorders. To alleviate this issue, various imputation methods for mass spectrometry-based omics data have been developed. However, a comprehensive comparison of these imputation methods is still lacking, and researchers are frequently confronted with a multitude of options without a clear rationale for method selection. To address this pressing need, we developed omicsMIC (mass spectrometry-based omics with Missing values Imputation methods Comparison platform), an interactive platform that provides researchers with a versatile framework to evaluate the performance of 28 diverse imputation methods. omicsMIC offers a nuanced perspective, acknowledging the inherent heterogeneity in biological data and the unique attributes of each dataset. Our platform empowers researchers to make data-driven decisions in imputation method selection based on real-time visualizations of the outcomes associated with different imputation strategies. The comprehensive benchmarking and versatility of omicsMIC make it a valuable tool for the scientific community engaged in mass spectrometry-based omics research. omicsMIC is freely available at https://github.com/WQLin8/omicsMIC.

Employing graph attention networks to decode psycho-metabolic interactions in Schizophrenia

Schizophrenia is a severe mental disorder characterized by intricate and underexplored interactions between psychological symptoms and metabolic health, presenting challenges in understanding the disease mechanisms and designing effective treatment strategies. To delve deeply into the complex interactions between mental and metabolic health in patients with schizophrenia, this study constructed a psycho-metabolic interaction network and optimized the Graph Attention Network (GAT). This approach reveals complex data patterns that traditional statistical analyses fail to capture. The results show that weight management and medication management play a central role in the interplay between psychiatric disorders and metabolic health. Furthermore, additional analysis revealed significant correlations between the history of psychiatric symptoms and physical health indicators, as well as the key roles of biochemical markers(e.g., triglycerides and low-density lipoprotein cholesterol), which have not been sufficiently emphasized in previous studies. This highlights the importance of medication management approaches, weight management, psychological treatment, and biomarker monitoring in comprehensive treatment and underscores the significance of the biopsychosocial model. This study is the first to utilize a GNN to explore the interactions between schizophrenia symptoms and metabolic features, providing new insights into understanding psychiatric disorders and guiding the development of more comprehensive treatment strategies for schizophrenia.

Improved detection and consistency of RNA-interacting proteomes using DIA SILAC

The RNA-interacting proteome is commonly characterized by UV-crosslinking followed by RNA purification, with protein recovery quantified using SILAC labeling followed by data-dependent acquisition (DDA) of proteomic data. However, the low efficiency of UV-crosslinking, combined with limited sensitivity of the DDA approach often restricts detection to relatively abundant proteins, necessitating multiple mass spec injections of fractionated peptides for each biological sample. Here we report an application of data-independent acquisition (DIA) with SILAC in a total RNA-associated protein purification (TRAPP) UV-crosslinking experiment. This gave 15% greater protein detection and lower inter-replicate variation relative to the same biological materials analyzed using DDA, while allowing single-shot analysis of the sample. As proof of concept, we determined the effects of arsenite treatment on the RNA-bound proteome of HEK293T cells. The DIA dataset yielded similar GO term enrichment for RNA-binding proteins involved in cellular stress responses to the DDA dataset while detecting extra proteins unseen by DDA. Overall, the DIA SILAC approach improved detection of proteins over conventional DDA SILAC for generating RNA-interactome datasets, at a lower cost due to reduced machine time. Analyses are described for TRAPP data, but the approach is suitable for proteomic analyses following essentially any RNA-binding protein enrichment technique.

Monotonic trends of soil microbiomes, metagenomic and metabolomic functioning across ecosystems along water gradients in the Altai region, northwestern China

To investigate microbial communities and their contributions to carbon and nutrient cycling along water gradients can enhance our comprehension of climate change impacts on ecosystem services. Thus, we conducted an assessment of microbial communities, metagenomic functions, and metabolomic profiles within four ecosystems, i.e., desert grassland (DG), shrub-steppe (SS), forest (FO), and marsh (MA) in the Altai region of Xinjiang, China. Our results showed that soil total carbon (TC), total nitrogen, NH4+, and NO3- increased, but pH decreased with soil water gradients. Microbial abundances and richness also increased with soil moisture except the abundances of fungi and protists being lowest in MA. A shift in microbial community composition is evident along the soil moisture gradient, with Proteobacteria, Basidiomycota, and Evosea proliferating but a decline in Actinobacteria and Cercozoa. The β-diversity of microbiomes, metagenomic, and metabolomic functioning were correlated with soil moisture gradients and have significant associations with specific soil factors of TC, NH4+, and pH. Metagenomic functions associated with carbohydrate and DNA metabolisms, as well as phages, prophages, TE, plasmids functions diminished with moisture, whereas the genes involved in nitrogen and potassium metabolism, along with certain biological interactions and environmental information processing functions, demonstrated an augmentation. Additionally, MA harbored the most abundant metabolomics dominated by lipids and lipid-like molecules and organic oxygen compounds, except certain metabolites showing decline trends along water gradients, such as N'-Hydroxymethylnorcotinine and 5-Hydroxyenterolactone. Thus, our study suggests that future ecosystem succession facilitated by changes in rainfall patterns will significantly alter soil microbial taxa, functional potential, and metabolite fractions.

Insights into serum metabolic biomarkers for early detection of incident diabetic kidney disease in Chinese patients with type 2 diabetes by random forest

Diabetic kidney disease (DKD) is a leading cause of end-stage renal disease (ESRD) worldwide. Early detection is critical for the risk stratification and early intervention of progressive DKD. Serum creatinine (sCr) and urine output are used to assess kidney function, but these markers are limited by their delayed changes following kidney pathology, and lacking of both sensitivity and accuracy. Hence, it is essential to illustrate potential diagnostic indicators to enhance the precise prediction of early DKD. A total of 194 Chinese individuals include 30 healthy participants (Stage 0) and 164 incidents with type 2 diabetes (T2D) spanning from DKD's Stage 1a to 4 were recruited and their serums were subjected for untargeted metabolomic analysis. Random forest (RF), a machine learning approach, together with univariate linear regression (ULR) and multivariate linear regression (MvLR) analysis were applied to characterize the features of untargeted metabolites of DKD patients and to identify candidate DKD biomarkers. Our results indicate that 2-(α-D-mannopyranosyl)-L-tryptophan (ADT), succinyladenosine (SAdo), pseudouridine and N,N,N-trimethyl-L-alanyl-L-proline betaine (L-L-TMAP) were associated with the development of DKD, in particular, the latter three that were significantly elevated in Stage 2-4 T2D incidents. Each of the four metabolites in combination with sCr achieves better performance than sCr alone with area under the receiver operating characteristic curve (AUC) of 0.81-0.91 in predicting DKD stages. An average of 3.9 years follow-up study of another cohort including 106 Stage 2-3 patients suggested that "urinary albumin-to-creatinine ratio (UACR) + ADT + SAdo" can be utilized for better prognosis evaluation of early DKD (average AUC = 0.9502) than UACR without sexual difference.

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.

Molecular fingerprint of gilthead seabream physiology in response to pollutant mixtures in the wild

The increase in trace element concentrations in the aquatic environment due to anthropogenic activities, urges the need for their monitoring and potential toxicity, persistence, bioaccumulation, and biomagnification at different trophic levels. Gilthead seabream is a species of commercial importance in the Mediterranean Sea, both for the aquaculture and fisheries sectors, however very little is known about their trace element contamination accumulation and the resulting effect on their health status. In the present study, 135 juveniles were collected from seven coastal lagoons known to be essential nursery areas for this species. We measured seventeen different inorganic contaminants at the individual level in fish muscle (namely Al, As, Be, Bi, Cd, Cr, Cu, Hg, Li, Ni, Pb, Rb, Sb, Sr, Ti, Tl and Zn). Our results revealed the accumulation of multiple trace elements in individuals and distinct contamination signatures between lagoons which might lead to contrasted quality as nurseries for juveniles of numerous ecologically and economically relevant fish species in addition to seabreams. We further evaluated the potential adverse effect of these complex contamination mixtures on the liver (the main organ implicated in the metabolism of xenobiotics) and red muscle (a highly metabolic organ) using a proteomic approach. Alterations in cellular organization pathways and protein transport were detected in both tissues (albeit they were not similarly regulated). Chromosome organization and telomere maintenance in the liver appeared to be affected by contaminant mixture which could increase mortality, age-related disease risk and shorter lifetime expectancy for these juveniles. Red muscle proteome also demonstrated an upregulation of pathways involved in metabolism in response to contamination which raises the issue of potential energy allocation trade-offs between the organisms' main functions such as reproduction and growth. This study provides new insights into the cellular and molecular responses of seabreams to environmental pollution and proposed biomarkers of health effects of trace elements that could serve as a starting point for larger-scale biomonitoring programs.

Deep Learning Methods for Omics Data Imputation

One common problem in omics data analysis is missing values, which can arise due to various reasons, such as poor tissue quality and insufficient sample volumes. Instead of discarding missing values and related data, imputation approaches offer an alternative means of handling missing data. However, the imputation of missing omics data is a non-trivial task. Difficulties mainly come from high dimensionality, non-linear or non-monotonic relationships within features, technical variations introduced by sampling methods, sample heterogeneity, and the non-random missingness mechanism. Several advanced imputation methods, including deep learning-based methods, have been proposed to address these challenges. Due to its capability of modeling complex patterns and relationships in large and high-dimensional datasets, many researchers have adopted deep learning models to impute missing omics data. This review provides a comprehensive overview of the currently available deep learning-based methods for omics imputation from the perspective of deep generative model architectures such as autoencoder, variational autoencoder, generative adversarial networks, and Transformer, with an emphasis on multi-omics data imputation. In addition, this review also discusses the opportunities that deep learning brings and the challenges that it might face in this field.

The Role of the Nuclear Receptor FXR in Arsenic-Induced Glucose Intolerance in Mice

Inorganic arsenic in drinking water is prioritized as a top environmental contaminant by the World Health Organization, with over 230 million people potentially being exposed. Arsenic toxicity has been well documented and is associated with a plethora of human diseases, including diabetes, as established in numerous animal and epidemiological studies. Our previous study revealed that arsenic exposure leads to the inhibition of nuclear receptors, including LXR/RXR. To this end, FXR is a nuclear receptor central to glucose and lipid metabolism. However, limited studies are available for understanding arsenic exposure-FXR interactions. Herein, we report that FXR knockout mice developed more profound glucose intolerance than wild-type mice upon arsenic exposure, supporting the regulatory role of FXR in arsenic-induced glucose intolerance. We further exposed mice to arsenic and tested if GW4064, a FXR agonist, could improve glucose intolerance and dysregulation of hepatic proteins and serum metabolites. Our data showed arsenic-induced glucose intolerance was remarkably diminished by GW4064, accompanied by a significant ratio of alleviation of dysregulation in hepatic proteins (83%) and annotated serum metabolites (58%). In particular, hepatic proteins "rescued" from arsenic toxicity by GW4064 featured members of glucose and lipid utilization. For instance, the expression of PCK1, a candidate gene for diabetes and obesity that facilitates gluconeogenesis, was repressed under arsenic exposure in the liver, but revived with the GW4064 supplement. Together, our comprehensive dataset indicates FXR plays a key role and may serve as a potential therapeutic for arsenic-induced metabolic disorders.

Development of a predictive nomogram for acute respiratory distress syndrome in patients with acute pancreatitis complicated with acute kidney injury

BACKGROUND

Acute respiratory distress syndrome (ARDS) is a common complication in patients with acute pancreatitis (AP), especially when patients complicated with acute kidney injury (AKI), resulting in increased duration of hospitalization and mortality. It is of potential clinical significance to develop a predictive model to identify the the high-risk patients.

METHOD

AP patients complicated with AKI from January 2019 to March 2022 were enrolled in this study and randomly divided into training cohort and validation cohort at a ratio of 2:1. The Least absolute shrinkage and selection operator(LASSO) regression and machine learning algorithms were applied to select features. A nomogram was developed based on the multivariate logistic regression. The performance of the nomogram was evaluated by AUC, calibration curves, and decision curve analysis.

RESULTS

A total of 292 patients were enrolled in the study, with 206 in the training cohort and 86 in the validation cohort. Multivariate logistic analysis showed that IAP (Odds Ratio (OR)=4.60, 95%CI:1.23-18.24, p = 0.02), shock (OR = 12.99, 95%CI:3.47-64.04, p < 0.001), CRP(OR= 26.19, 95%CI:9.37-85.57, p < 0.001), LDH (OR = 13.13, 95%CI:4.76-40.42, p < 0.001) were independent predictors of ARDS. The nomogram was developed based on IAP, shock, CRP and LDH. The nomogram showed good discriminative ability with an AUC value of 0.954 and 0.995 in the training and validation cohort, respectively. The calibration curve indicating good concordance between the predicted and observed values. The DCA showed favorable net clinical benefit.

CONCLUSION

This study developed a simple model for predicting ARDS in AP patients complicated with AKI. The nomogram can help clinicians identify high-risk patients and optimize therapeutic strategies.

omicsMIC: a Comprehensive Benchmarking Platform for Robust Comparison of Imputation Methods in Mass Spectrometry-based Omics Data

Mass spectrometry is a powerful and widely used tool for generating proteomics, lipidomics, and metabolomics profiles, which is pivotal for elucidating biological processes and identifying biomarkers. However, missing values in spectrometry-based omics data may pose a critical challenge for the comprehensive identification of biomarkers and elucidation of the biological processes underlying human complex disorders. To alleviate this issue, various imputation methods for mass spectrometry-based omics data have been developed. However, a comprehensive and systematic comparison of these imputation methods is still lacking, and researchers are frequently confronted with a multitude of options without a clear rationale for method selection. To address this pressing need, we developed omicsMIC (mass spectrometry-based omics with Missing values Imputation methods Comparison platform), an interactive platform that provides researchers with a versatile framework to simulate and evaluate the performance of 28 diverse imputation methods. omicsMIC offers a nuanced perspective, acknowledging the inherent heterogeneity in biological data and the unique attributes of each dataset. Our platform empowers researchers to make data-driven decisions in imputation method selection based on real-time visualizations of the outcomes associated with different imputation strategies. The comprehensive benchmarking and versatility of omicsMIC make it a valuable tool for the scientific community engaged in mass spectrometry-based omics research. OmicsMIC is freely available at https://github.com/WQLin8/omicsMIC.

An epidemiological introduction to human metabolomic investigations

Metabolomics holds great promise for uncovering insights around biological processes impacting disease in human epidemiological studies. Metabolites can be measured across biological samples, including plasma, serum, saliva, urine, stool, and whole organs and tissues, offering a means to characterize metabolic processes relevant to disease etiology and traits of interest. Metabolomic epidemiology studies face unique challenges, such as identifying metabolites from targeted and untargeted assays, defining standards for quality control, harmonizing results across platforms that often capture different metabolites, and developing statistical methods for high-dimensional and correlated metabolomic data. In this review, we introduce metabolomic epidemiology to the broader scientific community, discuss opportunities and challenges presented by these studies, and highlight emerging innovations that hold promise to uncover new biological insights.

How missing value imputation is confounded with batch effects and what you can do about it

In data-processing pipelines, upstream steps can influence downstream processes because of their sequential nature. Among these data-processing steps, batch effect (BE) correction (BEC) and missing value imputation (MVI) are crucial for ensuring data suitability for advanced modeling and reducing the likelihood of false discoveries. Although BEC-MVI interactions are not well studied, they are ultimately interdependent. Batch sensitization can improve the quality of MVI. Conversely, accounting for missingness also improves proper BE estimation in BEC. Here, we discuss how BEC and MVI are interconnected and interdependent. We show how batch sensitization can improve any MVI and bring attention to the idea of BE-associated missing values (BEAMs). Finally, we discuss how batch-class imbalance problems can be mitigated by borrowing ideas from machine learning.

The HuMet Repository: Watching human metabolism at work

The human metabolism constantly responds to stimuli such as food intake, fasting, exercise, and stress, triggering adaptive biochemical processes across multiple metabolic pathways. To understand the role of these processes and disruptions thereof in health and disease, detailed documentation of healthy metabolic responses is needed but still scarce on a time-resolved metabolome-wide level. Here, we present the HuMet Repository, a web-based resource for exploring dynamic metabolic responses to six physiological challenges (exercise, 36 h fasting, oral glucose and lipid loads, mixed meal, cold stress) in healthy subjects. For building this resource, we integrated existing and newly derived metabolomics data measured in blood, urine, and breath samples of 15 young healthy men at up to 56 time points during the six highly standardized challenge tests conducted over four days. The data comprise 1.1 million data points acquired on multiple platforms with temporal profiles of 2,656 metabolites from a broad range of biochemical pathways. By embedding the dataset into an interactive web application, we enable users to easily access, search, filter, analyze, and visualize the time-resolved metabolomic readouts and derived results. Users can put metabolites into their larger context by identifying metabolites with similar trajectories or by visualizing metabolites within holistic metabolic networks to pinpoint pathways of interest. In three showcases, we outline the value of the repository for gaining biological insights and generating hypotheses by analyzing the wash-out of dietary markers, the complementarity of metabolomics platforms in dynamic versus cross-sectional data, and similarities and differences in systemic metabolic responses across challenges. With its comprehensive collection of time-resolved metabolomics data, the HuMet Repository, freely accessible at https://humet.org/, is a reference for normal, healthy responses to metabolic challenges in young males. It will enable researchers with and without computational expertise, to flexibly query the data for their own research into the dynamics of human metabolism.

AlphaPeptStats: an open-source Python package for automated and scalable statistical analysis of mass spectrometry-based proteomics

SUMMARY

The widespread application of mass spectrometry (MS)-based proteomics in biomedical research increasingly requires robust, transparent, and streamlined solutions to extract statistically reliable insights. We have designed and implemented AlphaPeptStats, an inclusive Python package with currently with broad functionalities for normalization, imputation, visualization, and statistical analysis of label-free proteomics data. It modularly builds on the established stack of Python scientific libraries and is accompanied by a rigorous testing framework with 98% test coverage. It imports the output of a range of popular search engines. Data can be filtered and normalized according to user specifications. At its heart, AlphaPeptStats provides a wide range of robust statistical algorithms such as t-tests, analysis of variance, principal component analysis, hierarchical clustering, and multiple covariate analysis-all in an automatable manner. Data visualization capabilities include heat maps, volcano plots, and scatter plots in publication-ready format. AlphaPeptStats advances proteomic research through its robust tools that enable researchers to manually or automatically explore complex datasets to identify interesting patterns and outliers.

AVAILABILITY AND IMPLEMENTATION

AlphaPeptStats is implemented in Python and part of the AlphaPept framework. It is released under a permissive Apache license. The source code and one-click installers are freely available and on GitHub at https://github.com/MannLabs/alphapeptstats.

The BinDiscover database: a biology-focused meta-analysis tool for 156,000 GC-TOF MS metabolome samples

Metabolomics by gas chromatography/mass spectrometry (GC/MS) provides a standardized and reliable platform for understanding small molecule biology. Since 2005, the West Coast Metabolomics Center at the University of California at Davis has collated GC/MS metabolomics data from over 156,000 samples and 2000 studies into the standardized BinBase database. We believe that the observations from these samples will provide meaningful insight to biologists and that our data treatment and webtool will provide insight to others who seek to standardize disparate metabolomics studies. We here developed an easy-to-use query interface, BinDiscover, to enable intuitive, rapid hypothesis generation for biologists based on these metabolomic samples. BinDiscover creates observation summaries and graphics across a broad range of species, organs, diseases, and compounds. Throughout the components of BinDiscover, we emphasize the use of ontologies to aggregate large groups of samples based on the proximity of their metadata within these ontologies. This adjacency allows for the simultaneous exploration of entire categories such as "rodents", "digestive tract", or "amino acids". The ontologies are particularly relevant for BinDiscover's ontologically grouped differential analysis, which, like other components of BinDiscover, creates clear graphs and summary statistics across compounds and biological metadata. We exemplify BinDiscover's extensive applicability in three showcases across biological domains.

Tidyproteomics: an open-source R package and data object for quantitative proteomics post analysis and visualization

BACKGROUND

The analysis of mass spectrometry-based quantitative proteomics data can be challenging given the variety of established analysis platforms, the differences in reporting formats, and a general lack of approachable standardized post-processing analyses such as sample group statistics, quantitative variation and even data filtering. We developed tidyproteomics to facilitate basic analysis, improve data interoperability and potentially ease the integration of new processing algorithms, mainly through the use of a simplified data-object.

RESULTS

The R package tidyproteomics was developed as both a framework for standardizing quantitative proteomics data and a platform for analysis workflows, containing discrete functions that can be connected end-to-end, thus making it easier to define complex analyses by breaking them into small stepwise units. Additionally, as with any analysis workflow, choices made during analysis can have large impacts on the results and as such, tidyproteomics allows researchers to string each function together in any order, select from a variety of options and in some cases develop and incorporate custom algorithms.

CONCLUSIONS

Tidyproteomics aims to simplify data exploration from multiple platforms, provide control over individual functions and analysis order, and serve as a tool to assemble complex repeatable processing workflows in a logical flow. Datasets in tidyproteomics are easy to work with, have a structure that allows for biological annotations to be added, and come with a framework for developing additional analysis tools. The consistent data structure and accessible analysis and plotting tools also offers a way for researchers to save time on mundane data manipulation tasks.

Untargeted Metabolomics to Characterize the Urinary Chemical Landscape of E-Cigarette Users

The health and safety of using e-cigarette products (vaping) have been challenging to assess and further regulate due to their complexity. Inhaled e-cigarette aerosols contain chemicals with under-recognized toxicological profiles, which could influence endogenous processes once inhaled. We urgently need more understanding on the metabolic effects of e-cigarette exposure and how they compare to combustible cigarettes. To date, the metabolic landscape of inhaled e-cigarette aerosols, including chemicals originated from vaping and perturbed endogenous metabolites in vapers, is poorly characterized. To better understand the metabolic landscape and potential health consequences of vaping, we applied liquid chromatography-mass spectrometry (LC-MS) based nontargeted metabolomics to analyze compounds in the urine of vapers, cigarette smokers, and nonusers. Urine from vapers (n = 34), smokers (n = 38), and nonusers (n = 45) was collected for verified LC-HRMS nontargeted chemical analysis. The altered features (839, 396, and 426 when compared smoker and control, vaper and control, and smoker and vaper, respectively) among exposure groups were deciphered for their structural identities, chemical similarities, and biochemical relationships. Chemicals originating from e-cigarettes and altered endogenous metabolites were characterized. There were similar levels of nicotine biomarkers of exposure among vapers and smokers. Vapers had higher urinary levels of diethyl phthalate and flavoring agents (e.g., delta-decalactone). The metabolic profiles featured clusters of acylcarnitines and fatty acid derivatives. More consistent trends of elevated acylcarnitines and acylglycines in vapers were observed, which may suggest higher lipid peroxidation. Our approach in monitoring shifts of the urinary chemical landscape captured distinctive alterations resulting from vaping. Our results suggest similar nicotine metabolites in vapers and cigarette smokers. Acylcarnitines are biomarkers of inflammatory status and fatty acid oxidation, which were dysregulated in vapers. With higher lipid peroxidation, radical-forming flavoring, and higher level of specific nitrosamine, we observed a trend of elevated cancer-related biomarkers in vapers as well. Together, these data present a comprehensive profiling of urinary biochemicals that were dysregulated due to vaping.

Exploring the association of physical activity with the plasma and urine metabolome in adolescents and young adults

BACKGROUND

Regular physical activity elicits many health benefits. However, the underlying molecular mechanisms through which physical activity influences overall health are less understood. Untargeted metabolomics enables system-wide mapping of molecular perturbations which may lend insights into physiological responses to regular physical activity. In this study, we investigated the associations of habitual physical activity with plasma and urine metabolome in adolescents and young adults.

METHODS

This cross-sectional study included participants from the DONALD (DOrtmund Nutritional and Anthropometric Longitudinally Designed) study with plasma samples n = 365 (median age: 18.4 (18.1, 25.0) years, 58% females) and 24 h urine samples n = 215 (median age: 18.1 (17.1, 18.2) years, 51% females). Habitual physical activity was assessed using a validated Adolescent Physical Activity Recall Questionnaire. Plasma and urine metabolite concentrations were determined using ultra-high-performance liquid chromatography-tandem mass spectroscopy (UPLC-MS/MS) methods. In a sex-stratified analysis, we conducted principal component analysis (PCA) to reduce the dimensionality of metabolite data and to create metabolite patterns. Multivariable linear regression models were then applied to assess the associations between self-reported physical activity (metabolic equivalent of task (MET)-hours per week) with single metabolites and metabolite patterns, adjusted for potential confounders and controlling the false discovery rate (FDR) at 5% for each set of regressions.

RESULTS

Habitual physical activity was positively associated with the "lipid, amino acids and xenometabolite" pattern in the plasma samples of male participants only (β = 1.02; 95% CI: 1.01, 1.04, p = 0.001, adjusted p = 0.042). In both sexes, no association of physical activity with single metabolites in plasma and urine and metabolite patterns in urine was found (all adjusted p > 0.05).

CONCLUSIONS

Our explorative study suggests that habitual physical activity is associated with alterations of a group of metabolites reflected in the plasma metabolite pattern in males. These perturbations may lend insights into some of underlying mechanisms that modulate effects of physical activity.

Multi-omic integration via similarity network fusion to detect molecular subtypes of ageing

Molecular subtyping of brain tissue provides insights into the heterogeneity of common neurodegenerative conditions, such as Alzheimer's disease. However, existing subtyping studies have mostly focused on single data modalities and only those individuals with severe cognitive impairment. To address these gaps, we applied similarity network fusion, a method capable of integrating multiple high-dimensional multi-omic data modalities simultaneously, to an elderly sample spanning the full spectrum of cognitive ageing trajectories. We analyzed human frontal cortex brain samples characterized by five omic modalities: bulk RNA sequencing (18 629 genes), DNA methylation (53 932 CpG sites), histone acetylation (26 384 peaks), proteomics (7737 proteins) and metabolomics (654 metabolites). Similarity network fusion followed by spectral clustering was used for subtype detection, and subtype numbers were determined by Eigen-gap and rotation cost statistics. Normalized mutual information determined the relative contribution of each modality to the fused network. Subtypes were characterized by associations with 13 age-related neuropathologies and cognitive decline. Fusion of all five data modalities (n = 111) yielded two subtypes (n S1 = 53, n S2 = 58), which were nominally associated with diffuse amyloid plaques; however, this effect was not significant after correction for multiple testing. Histone acetylation (normalized mutual information = 0.38), DNA methylation (normalized mutual information = 0.18) and RNA abundance (normalized mutual information = 0.15) contributed most strongly to this network. Secondary analysis integrating only these three modalities in a larger subsample (n = 513) indicated support for both three- and five-subtype solutions, which had significant overlap, but showed varying degrees of internal stability and external validity. One subtype showed marked cognitive decline, which remained significant even after correcting for tests across both three- and five-subtype solutions (p Bonf = 5.9 × 10-3). Comparison to single-modality subtypes demonstrated that the three-modal subtypes were able to uniquely capture cognitive variability. Comprehensive sensitivity analyses explored influences of sample size and cluster number parameters. We identified highly integrative molecular subtypes of ageing derived from multiple high dimensional, multi-omic data modalities simultaneously. Fusing RNA abundance, DNA methylation, and histone acetylation measures generated subtypes that were associated with cognitive decline. This work highlights the potential value and challenges of multi-omic integration in unsupervised subtyping of post-mortem brain.

Does remaining teeth and dental prosthesis associate with social isolation? A six-year longitudinal study from the Japan Gerontological Evaluation Study (JAGES)

OBJECTIVES

Social isolation was associated with increased mortality and numerous adverse health outcomes. However, the longitudinal association between oral health and social isolation has not been studied. In this longitudinal prospective cohort study, the association between the number of remaining teeth and dental prosthesis use with social isolation after 6-years follow-up was examined.

METHODS

Functionally independent adults aged 65 years or older, who were not socially isolated in 2010, were followed up until 2016 in the Japan Gerontological Evaluation Study. Data from 26 417 participants were analysed after random forest imputation to address missing data. Logistic regression models were used to calculate the odds ratio (OR) for incident social isolation in 2016 after adjusting for age, sex, educational attainment, income, activities of daily living, living area and having depressive symptoms.

RESULTS

The mean age of the participants at baseline was 72.3 (SD = 5.0). A total of 1,127 (4.3%) participants were socially isolated at follow-up. Of these, 338 (3.2%) had ≥20 teeth (with or without using dental prosthesis), 171 (3.9%) had 10-19 teeth and used dental prosthesis, 112 (4.2%) had 10-19 teeth and did not use the dental prosthesis, 338 (5.1%) had 0-9 teeth and used dental prosthesis, and 168 (7.6%) had 0-9 teeth and did not use the dental prosthesis. Fully adjusted logistic regression models showed that the OR of incident social isolation was higher for those with fewer teeth; OR = 1.13 (95%CI = 0.96-1.33) for those with 10-19 teeth and OR = 1.36 (95%CI = 1.17-1.58) for those with 0-9 teeth, compared to those with ≥20 teeth. The OR of incident social isolation was lower for those who used a dental prosthesis [OR = 0.90, 95%CI = 0.80-1.02)] compared to those who did not use a dental prosthesis. The interaction between the number of teeth and dental prosthesis use demonstrated that the latter mitigated the incidence of social isolation for participants with tooth loss. Compared to those with ≥20 teeth (with or without prosthesis use), participants with 0-9 teeth that did not use a dental prosthesis were 79% [OR = 1.79, 95%CI = 1.49-2.19] more likely to be socially isolated, whereas participants with 0-9 teeth that used a dental prosthesis were only 23% [OR = 1.23, 95%CI = 1.05-1.45] more likely to be socially isolated.

CONCLUSION

Tooth loss was the main predictor for social isolation at follow-up, while no dental prostheses use was an additional risk factor. Dental prosthesis use may reduce the risk of social isolation especially in those with severe tooth loss.

MACHINE LEARNING MODELS FOR PREDICTING ACUTE KIDNEY INJURY IN PATIENTS WITH SEPSIS-ASSOCIATED ACUTE RESPIRATORY DISTRESS SYNDROME

ABSTRACT

Background: Acute kidney injury (AKI) is a prevalent and serious complication among patients with sepsis-associated acute respiratory distress syndrome (ARDS). Prompt and accurate prediction of AKI has an important role in timely intervention, ultimately improving the patients' survival rate. This study aimed to establish machine learning models to predict AKI via thorough analysis of data derived from electronic medical records. Method: The data of eligible patients were retrospectively collected from the Medical Information Mart for Intensive Care III database from 2001 to 2012. The primary outcome was the development of AKI within 48 hours after intensive care unit admission. Four different machine learning models were established based on logistic regression, support vector machine, random forest, and extreme gradient boosting (XGBoost). The performance of all predictive models was evaluated using the area under receiver operating characteristic curve, precision-recall curve, confusion matrix, and calibration plot. Moreover, the discrimination ability of the machine learning models was compared with that of the Sequential Organ Failure Assessment (SOFA) model. Results; Among 1,085 sepsis-associated ARDS patients included in this research, 375 patients (34.6%) developed AKI within 48 hours after intensive care unit admission. Twelve predictive variables were selected and further used to establish the machine learning models. The XGBoost model yielded the most accurate predictions with the highest area under receiver operating characteristic curve (0.86) and accuracy (0.81). In addition, a novel shiny application based on the XGBoost model was established to predict the probability of developing AKI among patients with sepsis-associated ARDS. Conclusions: Machine learning models could be used for predicting AKI in patients with sepsis-associated ARDS. Accordingly, a user-friendly shiny application based on the XGBoost model with reliable predictive performance was released online to predict the probability of developing AKI among patients with sepsis-associated ARDS.

Replication and mediation of the association between the metabolome and clinical markers of metabolic health in an adolescent cohort study

Metabolomics-derived metabolites (henceforth metabolites) may mediate the relationship between modifiable risk factors and clinical biomarkers of metabolic health (henceforth clinical biomarkers). We set out to study the associations of metabolites with clinical biomarkers and a potential mediation effect in a population of young adults. First, we conducted a systematic literature review searching for metabolites associated with 11 clinical biomarkers (inflammation markers, glucose, blood pressure or blood lipids). Second, we replicated the identified associations in a study population of n = 218 (88 males and 130 females, average age of 18 years) participants of the DONALD Study. Sex-stratified linear regression models adjusted for age and BMI and corrected for multiple testing were calculated. Third, we investigated our previously reported metabolites associated with anthropometric and dietary factors mediators in sex-stratified causal mediation analysis. For all steps, both urine and blood metabolites were considered. We found 41 metabolites in the literature associated with clinical biomarkers meeting our inclusion criteria. We were able to replicate an inverse association of betaine with CRP in women, between body mass index and C-reactive protein (CRP) and between body fat and leptin. There was no evidence of mediation by lifestyle-related metabolites after correction for multiple testing. We were only able to partially replicate previous findings in our age group and did not find evidence of mediation. The complex interactions between lifestyle factors, the metabolome, and clinical biomarkers warrant further investigation.

Binned Data Provide Better Imputation of Missing Time Series Data from Wearables

The presence of missing values in a time-series dataset is a very common and well-known problem. Various statistical and machine learning methods have been developed to overcome this problem, with the aim of filling in the missing values in the data. However, the performances of these methods vary widely, showing a high dependence on the type of data and correlations within the data. In our study, we performed some of the well-known imputation methods, such as expectation maximization, k-nearest neighbor, iterative imputer, random forest, and simple imputer, to impute missing data obtained from smart, wearable health trackers. In this manuscript, we proposed the use of data binning for imputation. We showed that the use of data binned around the missing time interval provides a better imputation than the use of a whole dataset. Imputation was performed for 15 min and 1 h of continuous missing data. We used a dataset with different bin sizes, such as 15 min, 30 min, 45 min, and 1 h, and we carried out evaluations using root mean square error (RMSE) values. We observed that the expectation maximization algorithm worked best for the use of binned data. This was followed by the simple imputer, iterative imputer, and k-nearest neighbor, whereas the random forest method had no effect on data binning during imputation. Moreover, the smallest bin sizes of 15 min and 1 h were observed to provide the lowest RMSE values for the majority of the time frames during the imputation of 15 min and 1 h of missing data, respectively. Although applicable to digital health data, we think that this method will also find applicability in other domains.

Alignment and Analysis of a Disparately Acquired Multibatch Metabolomics Study of Maternal Pregnancy Samples

Untargeted liquid chromatography-mass spectrometry metabolomics studies are typically performed under roughly identical experimental settings. Measurements acquired with different LC-MS protocols or following extended time intervals harbor significant variation in retention times and spectral abundances due to altered chromatographic, spectrometric, and other factors, raising many data analysis challenges. We developed a computational workflow for merging and harmonizing metabolomics data acquired under disparate LC-MS conditions. Plasma metabolite profiles were collected from two sets of maternal subjects three years apart using distinct instruments and LC-MS procedures. Metabolomics features were aligned using metabCombiner to generate lists of compounds detected across all experimental batches. We applied data set-specific normalization methods to remove interbatch and interexperimental variation in spectral intensities, enabling statistical analysis on the assembled data matrix. Bioinformatics analyses revealed large-scale metabolic changes in maternal plasma between the first and third trimesters of pregnancy and between maternal plasma and umbilical cord blood. We observed increases in steroid hormones and free fatty acids from the first trimester to term of gestation, along with decreases in amino acids coupled to increased levels in cord blood. This work demonstrates the viability of integrating nonidentically acquired LC-MS metabolomics data and its utility in unconventional metabolomics study designs.

Dealing with missing values in proteomics data

Proteomics data are often plagued with missingness issues. These missing values (MVs) threaten the integrity of subsequent statistical analyses by reduction of statistical power, introduction of bias, and failure to represent the true sample. Over the years, several categories of missing value imputation (MVI) methods have been developed and adapted for proteomics data. These MVI methods perform their tasks based on different prior assumptions (e.g., data is normally or independently distributed) and operating principles (e.g., the algorithm is built to address random missingness only), resulting in varying levels of performance even when dealing with the same dataset. Thus, to achieve a satisfactory outcome, a suitable MVI method must be selected. To guide decision making on suitable MVI method, we provide a decision chart which facilitates strategic considerations on datasets presenting different characteristics. We also bring attention to other issues that can impact proper MVI such as the presence of confounders (e.g., batch effects) which can influence MVI performance. Thus, these too, should be considered during or before MVI.

Machine Learning in Cardiac Surgery: Predicting Mortality and Readmission

Predicting outcomes in open-heart surgery can be challenging. Unexpected readmissions, long hospital stays, and mortality have economic implications. In this study, we investigated machine learning (ML) performance in data visualization and predicting patient outcomes associated with open-heart surgery. We evaluated 8,947 patients who underwent cardiac surgery from April 2006 to January 2018. Data visualization and classification were performed at cohort-level and patient-level using clustering, correlation matrix, and seven different predictive models for predicting three outcomes ("Discharged," "Died," and "Readmitted") at binary level. Cross-validation was used to train and test each dataset with the application of hyperparameter optimization and data imputation techniques. Machine learning showed promising performance for predicting mortality (AUC 0.83 ± 0.03) and readmission (AUC 0.75 ± 0.035). The cohort-level analysis revealed that ML performance is comparable to the Society of Thoracic Surgeons (STS) risk model even with limited number of samples ( e.g. , less than 3,000 samples for ML versus more than 100,000 samples for the STS risk models). With all cases (8,947 samples, referred as patient-level analysis), ML showed comparable performance to what has been reported for the STS models. However, we acknowledge that it remains unknown at this stage as to how the model might perform outside the institution and does not in any way constitute a comparison of the performance of the internal model with the STS model. Our study demonstrates a systematic application of ML in analyzing and predicting outcomes after open-heart surgery. The predictive utility of ML in cardiac surgery and clinical implications of the results are highlighted.

Forced vital capacity trajectories in patients with idiopathic pulmonary fibrosis: a secondary analysis of a multicentre, prospective, observational cohort

BACKGROUND

Idiopathic pulmonary fibrosis is a progressive fibrotic lung disease with a variable clinical trajectory. Decline in forced vital capacity (FVC) is the main indicator of progression; however, missingness prevents long-term analysis of patterns in lung function. We aimed to identify distinct clusters of lung function trajectory among patients with idiopathic pulmonary fibrosis using machine learning techniques.

METHODS

We did a secondary analysis of longitudinal data on FVC collected from a cohort of patients with idiopathic pulmonary fibrosis from the PROFILE study; a multicentre, prospective, observational cohort study. We evaluated the imputation performance of conventional and machine learning techniques to impute missing data and then analysed the fully imputed dataset by unsupervised clustering using self-organising maps. We compared anthropometric features, genomic associations, serum biomarkers, and clinical outcomes between clusters. We also performed a replication of the analysis on data from a cohort of patients with idiopathic pulmonary fibrosis from an independent dataset, obtained from the Chicago Consortium.

FINDINGS

415 (71%) of 581 participants recruited into the PROFILE study were eligible for further analysis. An unsupervised machine learning algorithm had the lowest imputation error among tested methods, and self-organising maps identified four distinct clusters (1-4), which was confirmed by sensitivity analysis. Cluster 1 comprised 140 (34%) participants and was associated with a disease trajectory showing a linear decline in FVC over 3 years. Cluster 2 comprised 100 (24%) participants and was associated with a trajectory showing an initial improvement in FVC before subsequently decreasing. Cluster 3 comprised 113 (27%) participants and was associated with a trajectory showing an initial decline in FVC before subsequent stabilisation. Cluster 4 comprised 62 (15%) participants and was associated with a trajectory showing stable lung function. Median survival was shortest in cluster 1 (2·87 years [IQR 2·29-3·40]) and cluster 3 (2·23 years [1·75-3·84]), followed by cluster 2 (4·74 years [3·96-5·73]), and was longest in cluster 4 (5·56 years [5·18-6·62]). Baseline FEV1 to FVC ratio and concentrations of the biomarker SP-D were significantly higher in clusters 1 and 3. Similar lung function clusters with some shared anthropometric features were identified in the replication cohort.

INTERPRETATION

Using a data-driven unsupervised approach, we identified four clusters of lung function trajectory with distinct clinical and biochemical features. Enriching or stratifying longitudinal spirometric data into clusters might optimise evaluation of intervention efficacy during clinical trials and patient management.

FUNDING

National Institute for Health and Care Research, Medical Research Council, and GlaxoSmithKline.

Metabolomics and modelling approaches for systems metabolic engineering

Metabolic engineering involves the manipulation of microbes to produce desirable compounds through genetic engineering or synthetic biology approaches. Metabolomics involves the quantitation of intracellular and extracellular metabolites, where mass spectrometry and nuclear magnetic resonance based analytical instrumentation are often used. Here, the experimental designs, sample preparations, metabolite quenching and extraction are essential to the quantitative metabolomics workflow. The resultant metabolomics data can then be used with computational modelling approaches, such as kinetic and constraint-based modelling, to better understand underlying mechanisms and bottlenecks in the synthesis of desired compounds, thereby accelerating research through systems metabolic engineering. Constraint-based models, such as genome scale models, have been used successfully to enhance the yield of desired compounds from engineered microbes, however, unlike kinetic or dynamic models, constraint-based models do not incorporate regulatory effects. Nevertheless, the lack of time-series metabolomic data generation has hindered the usefulness of dynamic models till today. In this review, we show that improvements in automation, dynamic real-time analysis and high throughput workflows can drive the generation of more quality data for dynamic models through time-series metabolomics data generation. Spatial metabolomics also has the potential to be used as a complementary approach to conventional metabolomics, as it provides information on the localization of metabolites. However, more effort must be undertaken to identify metabolites from spatial metabolomics data derived through imaging mass spectrometry, where machine learning approaches could prove useful. On the other hand, single-cell metabolomics has also seen rapid growth, where understanding cell-cell heterogeneity can provide more insights into efficient metabolic engineering of microbes. Moving forward, with potential improvements in automation, dynamic real-time analysis, high throughput workflows, and spatial metabolomics, more data can be produced and studied using machine learning algorithms, in conjunction with dynamic models, to generate qualitative and quantitative predictions to advance metabolic engineering efforts.

Nomogram-Based Chronic Kidney Disease Prediction Model for Type 1 Diabetes Mellitus Patients Using Routine Pathological Data

Type 1 diabetes mellitus (T1DM) patients are a significant threat to chronic kidney disease (CKD) development during their life. However, there is always a high chance of delay in CKD detection because CKD can be asymptomatic, and T1DM patients bypass traditional CKD tests during their routine checkups. This study aims to develop and validate a prediction model and nomogram of CKD in T1DM patients using readily available routine checkup data for early CKD detection. This research utilized 1375 T1DM patients’ sixteen years of longitudinal data from multi-center Epidemiology of Diabetes Interventions and Complications (EDIC) clinical trials conducted at 28 sites in the USA and Canada and considered 17 routinely available features. Three feature ranking algorithms, extreme gradient boosting (XGB), random forest (RF), and extremely randomized trees classifier (ERT), were applied to create three feature ranking lists, and logistic regression analyses were performed to develop CKD prediction models using these ranked feature lists to identify the best performing top-ranked features combination. Finally, the most significant features were selected to develop a multivariate logistic regression-based CKD prediction model for T1DM patients. This model was evaluated using sensitivity, specificity, accuracy, precision, and F1 score on train and test data. A nomogram of the final model was further generated for easy application in clinical practices. Hypertension, duration of diabetes, drinking habit, triglycerides, ACE inhibitors, low-density lipoprotein (LDL) cholesterol, age, and smoking habit were the top-8 features ranked by the XGB model and identified as the most important features for predicting CKD in T1DM patients. These eight features were selected to develop the final prediction model using multivariate logistic regression, which showed 90.04% and 88.59% accuracy in internal and test data validation. The proposed model showed excellent performance and can be used for CKD identification in T1DM patients during routine checkups.

The prognostic value of machine learning techniques versus cox regression model for head and neck cancer

BACKGROUND

Accurate prognostic prediction for head and neck cancer (HNC) is important for the improvement of clinical management. We aimed to compare the prognostic value of various machine learning techniques (MLTs) and statistical Cox regression model for different types of HNC.

METHODS

Clinical data of HNC patients were extracted from the Surveillance, Epidemiology, and End Results (SEER) database from 1974 to 2016. The prediction performance of five ML models, including random forest (RF), gradient boosting decision tree (GBDT), support vector machine (SVM), neural network (NN) and deep learning (DL), were compared with the statistical Cox regression model by estimating the concordance index (C-index), integrated Brier score (IBS), time-dependent receiver operating characteristic (ROC) curve and the area under the curve (AUC).

RESULTS

Our results showed that the RF model outperformed all other models in prognostic prediction for all tumor sites of HNC, particularly for major salivary gland cancer (MSGC, C-index: 88.730 ± 0.8700, IBS: 7.680 ± 0.4800), oral cavity cancer (OCC, C-index: 84.250 ± 0.6700, IBS: 11.480 ± 0.3300) and oropharyngeal cancer (OPC, C-index: 82.510 ± 0.5400, IBS: 10.120 ± 0.1400). Meanwhile, we analyzed the importance of each clinical variable in the RF model, in which age and tumor size presented the strongest positive prognostic effects. Additionally, similar results can be observed in the internal (6th edition of the AJCC TNM staging system cohort) and external validations (the TCGA HNC cohort).

CONCLUSIONS

The RF model is a promising prognostic prediction tool for HNC patients, regardless of the anatomic subsites.

Spectral binning as an approach to post-acquisition processing of high resolution FIE-MS metabolome fingerprinting data

INTRODUCTION

Flow infusion electrospray high resolution mass spectrometry (FIE-HRMS) fingerprinting produces complex, high dimensional data sets which require specialist in-silico software tools to process the data prior to analysis.

OBJECTIVES

Present spectral binning as a pragmatic approach to post-acquisition procession of FIE-HRMS metabolome fingerprinting data.

METHODS

A spectral binning approach was developed that included the elimination of single scan m/z events, the binning of spectra and the averaging of spectra across the infusion profile. The modal accurate m/z was then extracted for each bin. This approach was assessed using four different biological matrices and a mix of 31 known chemical standards analysed by FIE-HRMS using an Exactive Orbitrap. Bin purity and centrality metrics were developed to objectively assess the distribution and position of accurate m/z within an individual bin respectively.

RESULTS

The optimal spectral binning width was found to be 0.01 amu. 80.8% of the extracted accurate m/z matched to predicted ionisation products of the chemical standards mix were found to have an error of below 3 ppm. The open-source R package binneR was developed as a user friendly implementation of the approach. This was able to process 100 data files using 4 Central Processing Units (CPU) workers in only 55 seconds with a maximum memory usage of 1.36 GB.

CONCLUSION

Spectral binning is a fast and robust method for the post-acquisition processing of FIE-HRMS data. The open-source R package binneR allows users to efficiently process data from FIE-HRMS experiments with the resources available on a standard desktop computer.

Metabolic systems analysis identifies a novel mechanism contributing to shock in patients with endotheliopathy of trauma (EoT) involving thromboxane A2 and LTC4

Purpose

Endotheliopathy of trauma (EoT), as defined by circulating levels of syndecan-1 ≥ 40 ng/mL, has been reported to be associated with significantly increased transfusion requirements and a doubled 30-day mortality. Increased shedding of the glycocalyx points toward the endothelial cell membrane composition as important for the clinical outcome being the rationale for this study.

Results

The plasma metabolome of 95 severely injured trauma patients was investigated by mass spectrometry, and patients with EoT vs. non-EoT were compared by partial least square-discriminant analysis, identifying succinic acid as the top metabolite to differentiate EoT and non-EoT patients (VIP score = 3). EoT and non-EoT patients' metabolic flux profile was inferred by integrating the corresponding plasma metabolome data into a genome-scale metabolic network reconstruction analysis and performing a functional study of the metabolic capabilities of each group. Model predictions showed a decrease in cholesterol metabolism secondary to impaired mevalonate synthesis in EoT compared to non-EoT patients. Intracellular task analysis indicated decreased synthesis of thromboxanA2 and leukotrienes, as well as a lower carnitine palmitoyltransferase I activity in EoT compared to non-EoT patients. Sensitivity analysis also showed a significantly high dependence of eicosanoid-associated metabolic tasks on alpha-linolenic acid as unique to EoT patients.

Conclusions

Model-driven analysis of the endothelial cells' metabolism identified potential novel targets as impaired thromboxane A2 and leukotriene synthesis in EoT patients when compared to non-EoT patients. Reduced thromboxane A2 and leukotriene availability in the microvasculature impairs vasoconstriction ability and may thus contribute to shock in EoT patients. These findings are supported by extensive scientific literature; however, further investigations are required on these findings.

Identifying key risk factors for premature discontinuation of opioid use disorder treatment in the United States: A predictive modeling study

BACKGROUND

Treatment for opioid use disorder (OUD), particularly medication for OUD, is highly effective; however, retention in OUD treatment is a significant challenge. We aimed to identify key risk factors for premature exit from OUD treatment.

METHODS

We analyzed 2,381,902 cross-sectional treatment episodes for individuals in the U.S., discharged between Jan/1/2015 and Dec/31/2019. We developed classification models (Random Forest, Classification and Regression Trees (CART), Bagged CART, and Boosted CART), and analyzed 31 potential risk factors for premature treatment exit, including treatment characteristics, substance use history, socioeconomic status, and demographic characteristics. We stratified our analysis based on length of stay in treatment and service setting. Models were compared using cross-validation and the receiver operating characteristic area under the curve (ROC-AUC).

RESULTS

Random Forest outperformed other methods (ROC-AUC: 74%). The most influential risk factors included characteristics of service setting, geographic region, primary source of payment, and referral source. Race, ethnicity, and sex had far weaker predictive impacts. When stratified by treatment setting and length of stay, employment status and delay (days waited) to enter treatment were among the most influential factors. Their importance increased as treatment duration decreased. Notably, importance of referral source increased as the treatment duration increased. Finally, age and age of first use were important factors for lengths of stay of 2-7 days and in detox treatment settings.

CONCLUSIONS

The key factors of OUD treatment attrition identified in this analysis should be more closely explored (e.g., in causal studies) to inform targeted policies and interventions to improve models of care.

Addressing Missing Data in GC × GC Metabolomics: Identifying Missingness Type and Evaluating the Impact of Imputation Methods on Experimental Replication

Missing data is a significant issue in metabolomics that is often neglected when conducting data preprocessing, particularly when it comes to imputation. This can have serious implications for downstream statistical analyses and lead to misleading or uninterpretable inferences. In this study, we aim to identify the primary types of missingness that affect untargeted metabolomics data and compare strategies for imputation using two real-world comprehensive two-dimensional gas chromatography (GC × GC) data sets. We also present these goals in the context of experimental replication whereby imputation is conducted in a within-replicate-based fashion─the first description and evaluation of this strategy─and introduce an R package MetabImpute to carry out these analyses. Our results conclude that, in these two GC × GC data sets, missingness was most likely of the missing at-random (MAR) and missing not-at-random (MNAR) types as opposed to missing completely at-random (MCAR). Gibbs sampler imputation and Random Forest gave the best results when imputing MAR and MNAR compared against single-value imputation (zero, minimum, mean, median, and half-minimum) and other more sophisticated approaches (Bayesian principal component analysis and quantile regression imputation for left-censored data). When samples are replicated, within-replicate imputation approaches led to an increase in the reproducibility of peak quantification compared to imputation that ignores replication, suggesting that imputing with respect to replication may preserve potentially important features in downstream analyses for biomarker discovery.

Graph Properties of Mass-Difference Networks for Profiling and Discrimination in Untargeted Metabolomics

Untargeted metabolomics seeks to identify and quantify most metabolites in a biological system. In general, metabolomics results are represented by numerical matrices containing data that represent the intensities of the detected variables. These matrices are subsequently analyzed by methods that seek to extract significant biological information from the data. In mass spectrometry-based metabolomics, if mass is detected with sufficient accuracy, below 1 ppm, it is possible to derive mass-difference networks, which have spectral features as nodes and chemical changes as edges. These networks have previously been used as means to assist formula annotation and to rank the importance of chemical transformations. In this work, we propose a novel role for such networks in untargeted metabolomics data analysis: we demonstrate that their properties as graphs can also be used as signatures for metabolic profiling and class discrimination. For several benchmark examples, we computed six graph properties and we found that the degree profile was consistently the property that allowed for the best performance of several clustering and classification methods, reaching levels that are competitive with the performance using intensity data matrices and traditional pretreatment procedures. Furthermore, we propose two new metrics for the ranking of chemical transformations derived from network properties, which can be applied to sample comparison or clustering. These metrics illustrate how the graph properties of mass-difference networks can highlight the aspects of the information contained in data that are complementary to the information extracted from intensity-based data analysis.

Imputation of Missing Values for Multi-Biospecimen Metabolomics Studies: Bias and Effects on Statistical Validity

The analysis of high-throughput metabolomics mass spectrometry data across multiple biological sample types (biospecimens) poses challenges due to missing data. During differential abundance analysis, dropping samples with missing values can lead to severe loss of data as well as biased results in group comparisons and effect size estimates. However, the imputation of missing data (the process of replacing missing data with estimated values such as a mean) may compromise the inherent intra-subject correlation of a metabolite across multiple biospecimens from the same subject, which in turn may compromise the efficacy of the statistical analysis of differential metabolites in biomarker discovery. We investigated imputation strategies when considering multiple biospecimens from the same subject. We compared a novel, but simple, approach that consists of combining the two biospecimen data matrices (rows and columns of subjects and metabolites) and imputes the two biospecimen data matrices together to an approach that imputes each biospecimen data matrix separately. We then compared the bias in the estimation of the intra-subject multi-specimen correlation and its effects on the validity of statistical significance tests between two approaches. The combined approach to multi-biospecimen studies has not been evaluated previously even though it is intuitive and easy to implement. We examine these two approaches for five imputation methods: random forest, k nearest neighbor, expectation-maximization with bootstrap, quantile regression, and half the minimum observed value. Combining the biospecimen data matrices for imputation did not greatly increase efficacy in conserving the correlation structure or improving accuracy in the statistical conclusions for most of the methods examined. Random forest tended to outperform the other methods in all performance metrics, except specificity.

Assessment of label-free quantification and missing value imputation for proteomics in non-human primates

BACKGROUND

Reliable and effective label-free quantification (LFQ) analyses are dependent not only on the method of data acquisition in the mass spectrometer, but also on the downstream data processing, including software tools, query database, data normalization and imputation. In non-human primates (NHP), LFQ is challenging because the query databases for NHP are limited since the genomes of these species are not comprehensively annotated. This invariably results in limited discovery of proteins and associated Post Translational Modifications (PTMs) and a higher fraction of missing data points. While identification of fewer proteins and PTMs due to database limitations can negatively impact uncovering important and meaningful biological information, missing data also limits downstream analyses (e.g., multivariate analyses), decreases statistical power, biases statistical inference, and makes biological interpretation of the data more challenging. In this study we attempted to address both issues: first, we used the MetaMorphues proteomics search engine to counter the limits of NHP query databases and maximize the discovery of proteins and associated PTMs, and second, we evaluated different imputation methods for accurate data inference. We used a generic approach for missing data imputation analysis without distinguising the potential source of missing data (either non-assigned m/z or missing values across runs).

RESULTS

Using the MetaMorpheus proteomics search engine we obtained quantitative data for 1622 proteins and 10,634 peptides including 58 different PTMs (biological, metal and artifacts) across a diverse age range of NHP brain frontal cortex. However, among the 1622 proteins identified, only 293 proteins were quantified across all samples with no missing values, emphasizing the importance of implementing an accurate and statiscaly valid imputation method to fill in missing data. In our imputation analysis we demonstrate that Single Imputation methods that borrow information from correlated proteins such as Generalized Ridge Regression (GRR), Random Forest (RF), local least squares (LLS), and a Bayesian Principal Component Analysis methods (BPCA), are able to estimate missing protein abundance values with great accuracy.

CONCLUSIONS

Overall, this study offers a detailed comparative analysis of LFQ data generated in NHP and proposes strategies for improved LFQ in NHP proteomics data.