Protocol for constructing correlation-based molecular networks from large-scale untargeted metabolomics data.

Abstract

Here, we present a computational protocol for constructing correlation-based molecular networks from untargeted metabolomics data using MetVAE, a variational autoencoder-based framework. We describe steps for importing metabolomics features and sample metadata and for adjusting for compositionality, missingness, confounding, and high dimensionality. We then detail procedures for estimating sparse metabolite correlations and exporting GraphML files for network visualization in Cytoscape. By complementing spectral similarity networks, this protocol captures functional relationships reflected in cross-sample correlations.

Abstract as supplied by the indexing source. © the authors / publisher.