MCEE: a data preprocessing approach for metabolic confounding effect elimination
It is well recognized that physiological and environmental factors such as race, age, gender, and diurnal cycles often have a definite influence on metabolic results that statistically manifests as confounding variables. Currently, removal or controlling of confounding effects relies heavily on experimental design. There are no available data processing techniques focusing on the compensation of their effects. We therefore proposed a new method, Metabolic confounding effect elimination (MCEE), to remove the influence of specified confounding factors and make the data more accurate. The method consists of three steps: metabolites grouping, confounder-related metabolites selection, and metabolites modification. Its effectiveness and advantages were evaluated comprehensively by several simulated models and real datasets, and were compared with two typical methods, the principal component analysis (PCA)- and the direct orthogonal signal correction (DOSC)-based methods. MCEE is simple, effective, and safe, and is independent of sample number, association degree, and missing value. Hence, it may serve as a good complement to existing metabolomics data preprocessing methods and aid in better understanding the metabolic and biological status of interest.
KeywordsMetabolomics Confounding factor Generalized linear model Principal component analysis Direct orthogonal signal correction
This work was supported by the National Natural Science Foundation of China (31501079, 31500954 and 81772530), the National Key R&D Program of China (2017YFC0906800), and the Seventh Framework Programme of the European Union (294923). The authors thank the support of Biobank of Shanghai 6th People’s Hospital.
Compliance with ethical standards
The protocol of HCC was approved by the Zhongshan Hospital Institutional Review Board and written consents were signed by all participants before the study. The protocol of arthritis was approved by the Review Board in Institute of Basic Research in Clinical Medicine, China Academy of Chinese Medical Sciences, and all participants gave informed consent before they were involved in the study.
Conflict of Interest
The authors declare that they have no competing interests.
- 2.Hodson MP, Dear GJ, Roberts AD, Haylock CL, Ball RJ, Plumb RS, Stumpf CL, Griffin JL, Haselden JN. A gender-specific discriminator in Sprague-Dawley rat urine: the deployment of a metabolic profiling strategy for biomarker discovery and identification. Anal Biochem. 2007;362(2):182–92.CrossRefGoogle Scholar
- 5.Oberbach A, Bluher M, Wirth H, Till H, Kovacs P, Kullnick Y, Schlichting N, Tomm JM, Rolle-Kampczyk U, Murugaiyan J, Binder H, Dietrich A, von Bergen M. Combined proteomic and metabolomic profiling of serum reveals association of the complement system with obesity and identifies novel markers of body fat mass changes. J Proteome Res. 2011;10(10):4769–88.CrossRefGoogle Scholar
- 12.Pourhoseingholi MA, Baghestani AR, Vahedi M. How to control confounding effects by statistical analysis. Gastroenterol Hepatol Bed Bench. 2012;5(2):79–83.Google Scholar
- 18.Wang SY, Kuo CH, Tseng YJ. Batch Normalizer: a fast total abundance regression calibration method to simultaneously adjust batch and injection order effects in liquid chromatography/time-of-flight mass spectrometry-based metabolomics data and comparison with current calibration methods. Anal Chem. 2013;85(2):1037–46.CrossRefGoogle Scholar
- 33.Chen T, Cao Y, Zhang Y, Liu J, Bao Y, Wang C, Jia W, Zhao A. Random forest in clinical metabolomics for phenotypic discrimination and biomarker selection. Evid Based Complement Alternat Med. 2013;2013:298183.Google Scholar
- 35.Saag KG, Choi H (2006) Epidemiology, risk factors, and lifestyle modifications for gout. Arthritis Res Ther 8(Suppl 1:S2)Google Scholar
- 41.Fujii K, Tajiri K, Kajiwara T, Tanaka T, Murota K. Effects of NSAID on collagen and proteoglycan synthesis of cultured chondrocytes. J Rheumatol Suppl. 1989;18:28–31.Google Scholar
- 42.Palka J, Galewska Z. The effect of some antiinflammatory drugs on collagen of rat skin. Pol J Pharmacol Pharm. 1990;42(1):39–42.Google Scholar