Different statistical methods identify similar population-specific dietary patterns: an analysis of Longitudinal Study of Adult Health (ELSA-Brasil).

In recent decades, different data-driven approaches have emerged to identify dietary patterns (DP) and little is discussed about how these methods are able to capture diet complexity within the same population. This study aimed to apply three statistical methods to identify the DP of the Longitudinal Study of Adult Health (ELSA-Brasil) population and evaluate the similarities and differences between them. Dietary data were assessed at baseline in the ELSA-Brasil study using a FFQ. DP were identified by applying three statistical methods: (1) factor analysis (FA), (2) treelet transform (TT) and (3) reduced rank regression (RRR). The characteristics of individuals classified in the last tertile of each DP were compared. Cross-classification and Pearson's correlation coefficients were assessed to evaluate the agreement between individuals' adherence to DP of the three methods. A similar convenience DP was identified for all three methods. FA and TT also identified a similar prudent DP and a DP highly loaded for the food groups rice and beans. Individuals classified in the third tertile of similar DP of each method presented similar socio-demographic and nutrient intake characteristics. Regarding the cross-classification, prudent DP from FA and TT presented a higher level of agreement (75 %), while convenience DP from TT and RRR presented the lowest agreement (44·8 %). The different statistical methods were able to capture the populations' DP in a similar way while highlighting the particularities of each method.

Treelet transform analysis to identify clusters of systemic inflammatory variance in a population with moderate-to-severe traumatic brain injury.

BACKGROUND: Inflammatory cascades following traumatic brain injury (TBI) can have both beneficial and detrimental effects on recovery. Single biomarker studies do not adequately reflect the major arms of immunity and their relationships to long-term outcomes. Thus, we applied treelet transform (TT) analysis to identify clusters of interrelated inflammatory markers reflecting major components of systemic immune function for which substantial variation exists among individuals with moderate-to-severe TBI. METHODS: Serial blood samples from 221 adults with moderate-to-severe TBI were collected over 1-6 months post-injury (n = 607 samples). Samples were assayed for 33 inflammatory markers using Millipore multiplex technology. TT was applied to standardized mean biomarker values generated to identify latent patterns of correlated markers. Treelet clusters (TC) were characterized by biomarkers related to adaptive immunity (TC1), innate immunity (TC2), soluble molecules (TC3), allergy immunity (TC4), and chemokines (TC5). For each TC, a score was generated as the linear combination of standardized biomarker concentrations and cluster load for each individual in the cohort. Ordinal logistic or linear regression was used to test associations between TC scores and 6- and 12-month Glasgow Outcome Scale (GOS), Disability Rating Scale (DRS), and covariates. RESULTS: When adjusting for clinical covariates, TC5 was significantly associated with 6-month GOS (odds ratio, OR = 1.44; p-value, p = 0.025) and 6-month DRS scores (OR = 1.46; p = 0.013). TC5 relationships were attenuated when including all TC scores in the model (GOS: OR = 1.29, p = 0.163; DRS: OR = 1.33, p = 0.100). When adjusting for all TC scores and covariates, only TC3 was associated with 6- and 12-month GOS (OR = 1.32, p = 0.041; OR = 1.39, p = 0.002) and also 6- and 12-month DRS (OR = 1.38, p = 0.016; OR = 1.58, p = 0.0002). When applying TT to inflammation markers significantly associated with 6-month GOS, multivariate modeling confirmed that TC3 remained significantly associated with GOS. Biomarker cluster membership remained consistent between the GOS-specific dendrogram and overall dendrogram. CONCLUSIONS: TT effectively characterized chronic, systemic immunity among a cohort of individuals with moderate-to-severe TBI. We posit that chronic chemokine levels are effector molecules propagating cellular immune dysfunction, while chronic soluble receptors are inflammatory damage readouts perpetuated, in part, by persistent dysfunctional cellular immunity to impact neuro-recovery.

A review of statistical methods for dietary pattern analysis.

BACKGROUND: Dietary pattern analysis is a promising approach to understanding the complex relationship between diet and health. While many statistical methods exist, the literature predominantly focuses on classical methods such as dietary quality scores, principal component analysis, factor analysis, clustering analysis, and reduced rank regression. There are some emerging methods that have rarely or never been reviewed or discussed adequately. METHODS: This paper presents a landscape review of the existing statistical methods used to derive dietary patterns, especially the finite mixture model, treelet transform, data mining, least absolute shrinkage and selection operator and compositional data analysis, in terms of their underlying concepts, advantages and disadvantages, and available software and packages for implementation. RESULTS: While all statistical methods for dietary pattern analysis have unique features and serve distinct purposes, emerging methods warrant more attention. However, future research is needed to evaluate these emerging methods' performance in terms of reproducibility, validity, and ability to predict different outcomes. CONCLUSION: Selection of the most appropriate method mainly depends on the research questions. As an evolving subject, there is always scope for deriving dietary patterns through new analytic methodologies.

Dietary patterns generated by the Treelet Transform and risk of stroke: a Danish cohort study.

OBJECTIVE: To relate empirically derived dietary patterns identified using the Treelet Transform (TT) to risk of stroke. DESIGN: A prospective cohort study using the Danish Diet, Cancer and Health cohort. Dietary information was obtained in 1993-1997 using a validated semi-quantitative FFQ. Incident stroke diagnoses, obtained from the Danish National Patient Register, were verified by record review. Dietary patterns were generated using TT, and participants were categorised into quintiles based on their adherence to each pattern. Sex-specific Cox proportional hazard models estimated associations between dietary patterns and stroke. SETTING: Denmark. PARTICIPANTS: 55 061 men and women aged 50-64 years at the time of enrolment. RESULTS: Three dietary patterns explaining 15·4 % of the total variance were identified: a Prudent pattern, a Western pattern and a Wine & Snacks pattern. During a follow-up time of 10 years, 1513 cases occurred. Comparing the highest to lowest quintiles of intake, adherence to a Prudent pattern was inversely associated with stroke (HRmen 0·74, 95 % CI 0·60, 0·91; HRwomen 0·82, 95 % CI 0·62, 1·08), while adherence to a Western pattern was associated with greater risk (HRmen 1·61, 95 % CI 1·23, 2·10; HRwomen 2·01, 95 % CI 1·48, 2·72). No association was found for a Wine & Snacks pattern for women, but a weak inverse association was found for men (HR 0·81, 95 % CI 0·67, 0·99). CONCLUSIONS: The results of this study are broadly in line with current recommendations for a healthy diet to prevent stroke.

Patterns in metabolite profile are associated with risk of more aggressive prostate cancer: A prospective study of 3,057 matched case-control sets from EPIC.

Metabolomics may reveal novel insights into the etiology of prostate cancer, for which few risk factors are established. We investigated the association between patterns in baseline plasma metabolite profile and subsequent prostate cancer risk, using data from 3,057 matched case-control sets from the European Prospective Investigation into Cancer and Nutrition (EPIC). We measured 119 metabolite concentrations in plasma samples, collected on average 9.4 years before diagnosis, by mass spectrometry (AbsoluteIDQ p180 Kit, Biocrates Life Sciences AG). Metabolite patterns were identified using treelet transform, a statistical method for identification of groups of correlated metabolites. Associations of metabolite patterns with prostate cancer risk (OR1SD ) were estimated by conditional logistic regression. Supplementary analyses were conducted for metabolite patterns derived using principal component analysis and for individual metabolites. Men with metabolite profiles characterized by higher concentrations of either phosphatidylcholines or hydroxysphingomyelins (OR1SD = 0.77, 95% confidence interval 0.66-0.89), acylcarnitines C18:1 and C18:2, glutamate, ornithine and taurine (OR1SD = 0.72, 0.57-0.90), or lysophosphatidylcholines (OR1SD = 0.81, 0.69-0.95) had lower risk of advanced stage prostate cancer at diagnosis, with no evidence of heterogeneity by follow-up time. Similar associations were observed for the two former patterns with aggressive disease risk (the more aggressive subset of advanced stage), while the latter pattern was inversely related to risk of prostate cancer death (OR1SD = 0.77, 0.61-0.96). No associations were observed for prostate cancer overall or less aggressive tumor subtypes. In conclusion, metabolite patterns may be related to lower risk of more aggressive prostate tumors and prostate cancer death, and might be relevant to etiology of advanced stage prostate cancer.

Dietary Patterns Are Associated with Serum Metabolite Patterns and Their Association Is Influenced by Gut Bacteria among Older German Adults.

BACKGROUND: Although dietary intakes and dietary intake patterns (DPs) have been associated with single metabolites, it is unclear whether DPs are also reflected in specific metabolite patterns (MPs). Moreover, the influence of groups of gut bacteria on the relationship between DPs and MPs is underexplored. OBJECTIVES: We aimed to investigate the association of DPs and serum MPs and also the modifying effect of the gut bacteria compositional patterns (BCPs). METHODS: This is a cross-sectional investigation among 225 individuals (median age: 63 y; 53% women) from the European Prospective Investigation into Cancer and Nutrition study. Dietary intakes were assessed by three 24-h dietary recalls, gut bacteria composition was quantified by 16S rRNA gene sequencing, and the serum metabolome was profiled by an untargeted approach. We identified DPs and BCPs by the treelet transform analysis. We modeled associations between DPs and 8 previously published MPs and the modifying effect of BCPs by fitting generalized linear models using DataSHIELD R. RESULTS: We identified 5 DPs and 7 BCPs. The "bread, margarine, and processed meat" and "fruiting vegetables and vegetable oils" DPs were positively associated with the "amino acids" (ß = 0.35; 95% CI: 0.02, 0.69; P = 0.03) and "fatty acids" MPs (ß = 0.45; 95% CI: 0.16, 0.74; P = 0.01), respectively. The "tea and miscellaneous" was inversely associated with the "amino acids" (ß = -0.28; 95% CI: -0.52, -0.05; P = 0.02) and "amino acid derivatives" MPs (ß = -0.21; 95% CI: -0.39, -0.02; P = 0.03). One BCP negatively modified the association between the "bread, margarine, and processed meat" DP and the "amino acids" MP (P-interaction = 0.01). CONCLUSIONS: In older German adults, DPs are reflected in MPs, and the gut bacteria attenuate 1 DP-MP association. These MPs should be explored as biomarkers of these jointly consumed foods while taking into account a potentially modifying role of the gut bacteria.

Multi-Features Fusion for Fault Diagnosis of Pedal Robot Using Time-Speed Signals.

In order to realize automation of the pollutant emission tests of vehicles, a pedal robot is designed instead of a human-driven vehicle. Sometimes, the actual time-speed curve of the vehicle will deviate from the upper or lower limit of the worldwide light-duty test cycle (WLTC) target curve, which will cause a fault. In this paper, a new fault diagnosis method is proposed and applied to the pedal robot. Since principal component analysis (PCA), t-distributed stochastic neighbor embedding (t-SNE), and Autoencoder cannot extract feature information adequately when they are used alone, three types of feature components extracted by PCA, t-SNE, and Autoencoder are fused to form a nine-dimensional feature set. Then, the feature set is reduced into three-dimensional space via Treelet Transform. Finally, the fault samples are classified by Gaussian process classifier. Compared with the methods using only one algorithm to extract features, the proposed method has the minimum standard deviation, 0.0078, and almost the maximum accuracy, 98.17%. The accuracy of the proposed method is only 0.24% lower than that without Treelet Transform, but the processing time is 6.73% less than that without Treelet Transform. These indicate that the multi-features fusion model and Treelet Transform method is quite effective. Therefore, the proposed method is quite helpful for fault diagnosis of the pedal robot.

A treelet transform analysis to relate nutrient patterns to the risk of hormonal receptor-defined breast cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC).

OBJECTIVE: Pattern analysis has emerged as a tool to depict the role of multiple nutrients/foods in relation to health outcomes. The present study aimed at extracting nutrient patterns with respect to breast cancer (BC) aetiology. DESIGN: Nutrient patterns were derived with treelet transform (TT) and related to BC risk. TT was applied to twenty-three log-transformed nutrient densities from dietary questionnaires. Hazard ratios (HR) and 95 % confidence intervals computed using Cox proportional hazards models quantified the association between quintiles of nutrient pattern scores and risk of overall BC, and by hormonal receptor and menopausal status. Principal component analysis was applied for comparison. SETTING: The European Prospective Investigation into Cancer and Nutrition (EPIC). SUBJECTS: Women (n 334 850) from the EPIC study. RESULTS: The first TT component (TC1) highlighted a pattern rich in nutrients found in animal foods loading on cholesterol, protein, retinol, vitamins B12 and D, while the second TT component (TC2) reflected a diet rich in ß-carotene, riboflavin, thiamin, vitamins C and B6, fibre, Fe, Ca, K, Mg, P and folate. While TC1 was not associated with BC risk, TC2 was inversely associated with BC risk overall (HRQ5 v. Q1=0·89, 95 % CI 0·83, 0·95, P trend<0·01) and showed a significantly lower risk in oestrogen receptor-positive (HRQ5 v. Q1=0·89, 95 % CI 0·81, 0·98, P trend=0·02) and progesterone receptor-positive tumours (HRQ5 v. Q1=0·87, 95 % CI 0·77, 0·98, P trend<0·01). CONCLUSIONS: TT produces readily interpretable sparse components explaining similar amounts of variation as principal component analysis. Our results suggest that participants with a nutrient pattern high in micronutrients found in vegetables, fruits and cereals had a lower risk of BC.

Serum Lipid and Serum Metabolite Components in relation to anthropometric parameters in EPIC-Potsdam participants.

BACKGROUND/AIM: Lipidomic and metabolomic techniques become more and more important in human health research. Recent developments in analytical techniques enable the investigation of high amounts of substances. The high numbers of metabolites and lipids that are detected with among others mass spectrometric techniques challenge in most cases the statistical processes to bring out stable and interpretable results. This study targets to use the novel non-established statistical method treelet transform (TT) to investigate high numbers of metabolites and lipids and to compare the results with the established method principal component analysis (PCA). Serum lipid and metabolite profiles are investigated regarding their association to anthropometric parameters associated to obesity. METHODS: From 226 participants of the EPIC (European Prospective Investigation into Cancer and Nutrition)-Potsdam study blood samples were investigated with an untargeted metabolomics approach regarding serum metabolites and lipids. Additionally, participants were surveyed anthropometrically to assess parameters of obesity, such as body mass index (BMI), waist-to-hip-ratio (WHR) and body fat mass. TT and PCA are used to generate treelet components (TCs) and factors summarizing serum metabolites and lipids in new, latent variables without too much loss of information. With partial correlations TCs and factors were associated to anthropometry under the control for relevant parameters, such as sex and age. RESULTS: TT with metabolite variables (p=121) resulted in 5 stable and interpretable TCs explaining 18.9% of the variance within the data. PCA on the same variables generated 4 quite complex, less easily interpretable factors explaining 37.5% of the variance. TT on lipidomic data (p=353) produced 3 TCs as well as PCA on the same data resulted in 3 factors; the proportion of explained variance was 17.8% for TT and 39.8% for PCA. In both investigations TT ended up with stable components that are easier to interpret than the factors from the PCA. In general, the generated TCs and factors were similar in their structure when the factors are considered regarding the original variables loading high on them. Both TCs and factors showed associations to anthropometric measures. CONCLUSIONS: TT is a suitable statistical method to generate summarizing, latent variables in data sets with more variables than observations. In the present investigation it resulted in similar latent variables compared to the established method of PCA. Whereby less variance is explained by the summarizing constructs of TT compared to the factors of PCA, TCs are easier to interpret. Additionally the resulting TCs are quite stable in bootstrap samples.