NMR Metabolomics in Serum Fingerprinting of Schizophrenia Patients in a Serbian Cohort

Abstract

Schizophrenia is a widespread mental disorder that leads to significant functional impairments and premature death. The state of the art indicates gaps in the understanding and diagnosis of this disease, but also the need for personalized and precise approaches to patients through customized medical treatment and reliable monitoring of treatment response. In order to fulfill existing gaps, the establishment of a universal set of disorder biomarkers is a necessary step. Metabolomic investigations of serum samples of Serbian patients with schizophrenia (51) and healthy controls (39), based on NMR analyses associated with chemometrics, led to the identification of 26 metabolites/biomarkers for this disorder. Principal component analysis (PCA) and orthogonal partial least squares discriminant analysis (OPLS-DA) models with prediction accuracies of 0.9718 and higher were accomplished during chemometric analysis. The established biomarker set includes aspartate/aspartic acid, lysine, 2-hydroxybutyric acid, and acylglycerols, which are identified for the first time in schizophrenia serum samples by NMR experiments. The other 22 identified metabolites in the Serbian samples are in accordance with the previously established NMR-based serum biomarker sets of Brazilian and/or Chinese patient samples. Thirteen metabolites (lactate/lactic acid, threonine, leucine, isoleucine, valine, glutamine, asparagine, alanine, gamma-aminobutyric acid, choline, glucose, glycine and tyrosine) that are common for three different ethnic and geographic origins (Serbia, Brazil and China) could be a good start point for the setup of a universal NMR serum biomarker set for schizophrenia.

Metabolomic Profiling of Bipolar Disorder by 1H-NMR in Serbian Patients

“…Furthermore, the same dataset object also includes accompanying categorical and class variables (sample labels, axis scale, response class variable, triplicate grouping variable, cross-validation index variable, etc.) necessary for further chemometric analysis [ 35 ]. As a result, the data set table comprises 116 samples organized in triplicates from 39 patients belonging to the ‘Control’ healthy group and 102 samples of triplicates from 34 patients of the ‘BD’ group of patients, constituting 218 samples overall.…”

“…As a result of the splitting procedure, the training data set consists of 143 total samples: 77 samples from 26 individuals of ‘Control’ group and 66 samples from 22 patients of the ‘BD’ group, whereas the test data set includes 71 samples among which 33 samples from 11 patients of the ‘BD’ group and 38 samples from 13 ‘Control’ group individuals. All samples in both data sets are grouped as triplicates for each certain patient, and their allocation inside the data set was assigned to a specific categorical variable [ 35 ].…”

1H NMR‐based process understanding and biochemical marker identification methodology for monitoring CHO cell culture process during commercial‐scale manufacturing

“…The advantage of OPLS‐DA used in the identification of sample variations, is that a single component is used as a predictor for the class, while the other components describe the variation orthogonal to the first predictive component, which has benefits in exploring variables that contribute more to the separation of groups. The OPLS‐DA model has been proven to be an effective and reasonable method in the identification of differential metabolites and biomarkers in various scientific scopes [ 27–31 ] . To avoid the noises negatively affecting the signal distribution, binned spectra were furtherly aligned, mean‐centered and scaled to unit variance before the OPLS‐DA modeling.…”

“…The OPLS-DA model has been proven to be an effective and reasonable method in the identification of differential metabolites and biomarkers in various scientific scopes. [27][28][29][30][31] To avoid the noises negatively affecting the signal distribution, binned spectra were furtherly aligned, mean-centered and scaled to unit variance before the OPLS-DA modeling. Then, the S-line plot would be derived, which is tailor-made visualization and interpretation tool for NMR spectroscopy data.…”

Exploring the role of ketone bodies in the diagnosis and treatment of psychiatric disorders

“…Metabolomics studies have shown that the overall metabolic profiles of schizophrenia patients are demonstrably altered and differentiable from controls (193). Early proof-of-concept studies have demonstrated that schizophrenic patients can be accurately identified based on profiles of as few as four metabolomic biomarkers alone (194)(195)(196), which represents a compelling new perspective on diagnosis possibilities. Most recently, 86 differential metabolites and over 180 differential expression genes have been identified in a recent transcriptomic analysis study (197), with pathway-based integration indicating that the synthesis and degradation of KBs appears to be closely connected to schizophrenia pathogenesis.…”