Non-targeted UHPLC-MS metabolomic data processing methods: a comparative investigation of normalisation, missing value imputation, transformation and scaling
Методы обработки метаболомных данных UHPLC-MS без целевой направленности: сравнительное исследование нормализации, заполнения пропусков, преобразования и масштабирования
2016-04-14
SCID: 54.1/vwzzjssg
Discuss with AI
PCA and PLS-DA preprocessing recommendationsPQN normalizationUHPLC-MS metabolomic data processingglog (generalised logarithm) transformationmissing value imputation (Random Forest, KNN, none)
Figures from the paper
Abstract (AI)
INTRODUCTION: The generic metabolomics data processing workflow is constructed with a serial set of processes including peak picking, quality assurance, normalisation, missing value imputation, transformation and scaling. The combination of these processes should present the experimental data in an appropriate structure so to identify the biological changes in a valid and robust manner. OBJECTIVES: Currently, different researchers apply different data processing methods and no assessment of the permutations applied to UHPLC-MS datasets has been published. Here we wish to define the most appropriate data processing workflow. METHODS: We assess the influence of normalisation, missing value imputation, transformation and scaling methods on univariate and multivariate analysis of UHPLC-MS datasets acquired for different mammalian samples. RESULTS: , retention time or response. Following an exhaustive evaluation, we recommend PQN normalisation with no missing value imputation and no transformation or scaling for univariate analysis. For PCA we recommend applying PQN normalisation with Random Forest missing value imputation, glog transformation and no scaling method. For PLS-DA we recommend PQN normalisation, KNN as the missing value imputation method, generalised logarithm transformation and no scaling. These recommendations are based on searching for the biologically important metabolite features independent of their measured abundance. CONCLUSION: The appropriate choice of normalisation, missing value imputation, transformation and scaling methods differs depending on the data analysis method and the choice of method is essential to maximise the biological derivations from UHPLC-MS datasets.
Key Findings
1
For PCA, PQN normalisation combined with Random Forest missing value imputation, generalized log (glog) transformation, and no scaling is recommended.
2
For PLS-DA, PQN normalisation with KNN missing value imputation, generalized logarithm transformation, and no scaling is recommended.
3
For univariate analysis, Probabilistic Quotient Normalisation (PQN) with no missing value imputation and no transformation or scaling is recommended.
4
Method-specific preprocessing choices should prioritize finding biologically important metabolite features independent of measured abundance.
5
The combination of normalisation, missing value imputation, transformation and scaling critically affects univariate and multivariate analysis of UHPLC-MS metabolomic data.
Research Object
Non-targeted UHPLC-MS metabolomic datasets from mammalian samples
Research Subject
Effects of different data processing methods (normalisation, missing value imputation, transformation, and scaling) on univariate and multivariate analysis and identification of biologically important metabolite features
Publication Details
Publication Date
2016-04-14
Journal
Publisher
ISSN
Cited by
429
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest