Evaluation of Single Sample Network Inference Methods for Metabolomics-Based Systems Medicine

Abstract

Networks and network analyses are fundamental tools of systems biology. Networks are built by inferring pair-wise relationships among biological entities from a large number of samples such that subject-specific information is lost. The possibility of constructing these sample (individual)-specific networks from single molecular profiles might offer new insights in systems and personalized medicine and as a consequence is attracting more and more research interest. In this study, we evaluated and compared LIONESS (Linear Interpolation to Obtain Network Estimates for Single Samples) and ssPCC (single sample network based on Pearson correlation) in the metabolomics context of metabolite–metabolite association networks. We illustrated and explored the characteristics of these two methods on (i) simulated data, (ii) data generated from a dynamic metabolic model to simulate real-life observed metabolite concentration profiles, and (iii) 22 metabolomic data sets and (iv) we applied single sample network inference to a study case pertaining to the investigation of necrotizing soft tissue infections to show how these methods can be applied in metabolomics. We also proposed some adaptations of the methods that can be used for data exploration. Overall, despite some limitations, we found single sample networks to be a promising tool for the analysis of metabolomics data.

Exploration of Blood Lipoprotein and Lipid Fraction Profiles in Healthy Subjects through Integrated Univariate, Multivariate, and Network Analysis Reveals Association of Lipase Activity and Cholesterol Esterification with Sex and Age

“…There is a wealth of information contained in the relationships among plasma and blood metabolites [ 28 , 30 , 31 ], which are better captured using correlation measures as an index of association [ 32 , 33 ]. Lipoprotein main fraction association networks were built using the PCLRC algorithm and a Gaussian Graphical Model approach to estimate the pairwise partial correlations among the concentrations of the lipoprotein fractions.…”

“…Since lipoprotein concentrations change in an orchestrated fashion, the patterns of associations between lipoprotein fractions can be considered, to some extent, related to the underlying structure of the biological networks [ 20 ]. Differences in lipoprotein associations which are sex- and age-related can indeed point to affected molecular mechanisms since changes can be more significant than levels alone [ 21 , 22 ], as shown in applications to health, sex, and age phenotyping [ 17 , 23 ], cardiovascular risk [ 24 , 25 , 26 ], and bacterial infections [ 27 , 28 ].…”

Defining Blood Plasma and Serum Metabolome by GC-MS

“…In addition to standard approaches for statistical modeling, enrichment analysis [194] and network inference [195] methods are used to help integrate metabolomics data into multi-omics models. The development of deep learning (DL) technologies is anticipated in metabolomics.…”

Comparative Analyses of Gene Co-expression Networks: Implementations and Applications in the Study of Evolution

“…Therefore,a comparative method to uniformly analyze cross-condition or cross-species gene expression data is essential. Further, the potential construction and comparison of samplespecific GCNs from single transcriptomic profiles may offer new insights into network evolution and better understand samplespecific differences (Kuijjer et al, 2019;Jahagirdar and Saccenti, 2020).…”