Supplementary Information Influence of Missing Values Substitutes on Multivariate Analysis of Metabolomics Data 1. Principal Components Analysis (PCA)

As discussed in the main text after PCA, the original data are reduced to a number of significant PCs. Figure S1A displays the PCA procedure where the original data are converted into significant PCs by a decomposition of the data matrix X into two matrices, where the series of columns are represented as scores, T, and the series of rows are termed loadings, P. Additionally, both are orthogonal. The loadings can be used to understand which input variables are important as the weight for each original variable when computing the PC. Moreover, the first vectors demonstrating both series (scores and loadings) are called then eigenvectors of the first PC, the second vectors are named the eigenvectors of second PC and so on [1,2]. Figure S1. (A) Matrix structure for PCA, where loadings summarise variation in variables and scores summarise variation in samples. This results in an approximated matrix with reduced noise (T = P × X) [1]; (B) A graphical illustration of multiblock PCA. 2. Multiblock PCA Figure S1A displays the scheme for the PCA method, which is presented here for comparing with the multiblock extension of the method. Where, a data matrix with descriptor X can be considered here as a typical output of GC-MS with variables representing different metabolites. In contrast to PCA as shown in Figure S1B multiblock PCA associates three blocks of descriptors, the variables (metabolites) measured for samples that correspond to three cases normoxia, hypoxia and anoxia as described in manuscript, where the consensus direction among all the of them is pursued. As the method is closely relate to PCA, hence, latent variable and scores are calculated for each of the blocks, in the next step of calculation these values are combined into one super score block. The final output displays the consensus of all projections comprised in super scores Tsup, where weight wʹt highlight the significanc