Critical Assessment of the Biomarker Discovery and Classification Methods for Multiclass Metabolomics

Abstract

Multiclass metabolomics has been widely applied in clinical practice to understand pathophysiological processes involved in disease progression and diagnostic biomarkers of various disorders. In contrast to the binary problem, the multiclass classification problem is more difficult in terms of obtaining reliable and stable results due to the increase in the complexity of determining exact class decision boundaries. In particular, methods of biomarker discovery and classification have a significant effect on the multiclass model because different methods with significantly varied theories produce conflicting results even for the same dataset. However, a systematic assessment for selecting the most appropriate methods of biomarker discovery and classification for multiclass metabolomics is still lacking. Therefore, a comprehensive assessment is essential to measure the suitability of methods in multiclass classification models from multiple perspectives. In this study, five biomarker discovery methods and nine classification methods were assessed based on four benchmark datasets of multiclass metabolomics. The performance assessment of the biomarker discovery and classification methods was performed using three evaluation criteria: assessment a (cluster analysis of sample grouping), assessment b (biomarker consistency in multiple subgroups), and assessment c (accuracy in the classification model). As a result, 13 combining strategies with superior performance were selected under multiple criteria based on these benchmark datasets. In conclusion, superior strategies that performed consistently well are suggested for the discovery of biomarkers and the construction of a classification model for multiclass metabolomics.

Strategy for Identifying a Robust Metabolomic Signature Reveals the Altered Lipid Metabolism in Pituitary Adenoma

“…For each sample group, 100 training-validation sets were used to generate a metabolomic signature based on a consensus evaluation-based feature selection algorithm. This algorithm consists of two parts, namely, SVM-RFE and feature elimination based on ranking consensus evaluation (FERCE), the latter of which is a novel algorithm proposed in this study. SVM-RFE has a successful application in the biomarker identification of metabolomic data, particularly in addressing challenges such as noisy training examples, a low number of biological replicates, and nonlinearity among metabolites …”

Wekemo Bioincloud: A user‐friendly platform for meta‐omics data analyses

“…MetaProteomeAnalyzer [ 18 ] is a workflow for metaproteomic data analyses. Meanwhile, numerous innovative approaches have emerged for identifying reliable and stable biomarkers from ‐omics data [ 19 , 20 , 21 , 22 ], and several research have diligently summarized and compared various R packages or software tools designed for ‐omics data [ 23 , 24 , 25 , 26 , 27 , 28 , 29 , 30 , 31 ]. However, the majority of these tools are oriented toward one or two specific types of ‐omics data analyses.…”

MultiClassMetabo: A Superior Classification Model Constructed Using Metabolic Markers in Multiclass Metabolomics

“…To date, various methods have been used for constructing classification models in multiclass metabolomics, such as random forests. , Due to the differences in the principles of these classification methods, conflicting outcomes are observed when different classification methods are applied, even for the same data set. , Therefore, it is highly necessary to apply an appropriate classification method with superior performance for a specific data set. To select the most suitable method, the performance of all classification methods must be assessed using well-established criteria. , Apart from classification methods, methods of identifying metabolic markers are also of importance for the performance of the classification model in multiclass metabolomics. , …”