Skip to main content
Fig. 2 | BMC Bioinformatics

Fig. 2

From: Tensor decomposition-based and principal-component-analysis-based unsupervised feature extraction applied to the gene expression and methylation profiles in the brains of social insects with multiple castes

Fig. 2

A flow chart of PCA- and TD-based unsupervised FE. Gene expression and methylation profiles were examined by PCA or TD. For PCA, gene expression and methylation profiles were processed separately, whereas TD was applied after generating a tensor from them. For PCA, principal component (PC) loading attributed to samples was studied and selected for FE. Because PC loadings and PC scores attributed to genes show one-to-one correspondence, PC scores corresponding to the selected PC loadings were subjected to FE. For TD, one-sample singular value vectors used for FE were selected. Afterwards, during analysis of core tensors, G, gene singular value vectors associated with Gs with larger absolute values were selected. By means of the identified PC scores or gene singular value vectors, P-values were determined for genes, assuming a χ2 distribution, and genes associated with adjusted P-values less than 0.01 were finally selected

Back to article page