Multi-omics to predict changes during cold pressor test

Abstract

Background: The cold pressor test (CPT) is a widely used pain provocation test to investigate both pain tolerance and cardiovascular responses. We hypothesize, that performing multi-omic analyses during CPT gives the opportunity to home in on molecular mechanisms involved. Twenty-two females were phenotypically assessed before and after a CPT, and blood samples were taken. RNA-Sequencing, steroid profiling and untargeted metabolomics were performed. Each 'omic level was analyzed separately at both single-feature and systems-level (principal component [PCA] and partial least squares [PLS] regression analysis) and all 'omic levels were combined using an integrative multi-omics approach, all using the paired-sample design.

Results: We showed that PCA was not able to discriminate time points, while PLS did significantly distinguish time points using metabolomics and/or transcriptomic data, but not using conventional physiological measures. Transcriptomic and metabolomic data revealed at feature-, systems- and integrative- level biologically relevant processes involved during CPT, e.g. lipid metabolism and stress response.

Conclusion: Multi-omics strategies have a great potential in pain research, both at feature- and systems- level. Therefore, they should be exploited in intervention studies, such as pain provocation tests, to gain knowledge on the biological mechanisms involved in complex traits.

Keywords: Cold pressor test; Data integration; Metabolomics; Multi-omics; PCA; PLS; Systems biology; Transcriptomics.

Fig. 1

Clustering of the ‘omics samples before (blue circles) and after (orange triangles) the cold pressor test using A) multilevel PCA (unsupervised clustering) and B) multilevel sPLS-DA (supervised clustering). No significant prediction between the time points could be made using PCA, though using sPLS-DA both transcriptomics and metabolomics enabled significant prediction of before vs. after cold pressor test

Fig. 2

Network of correlating genes and metabolites, detected by DIABLO. Genes in blue, metabolite IDs in red

Fig. 3

Two of the top differentially expressed metabolites (A) and genes (B) before vs. after cold pressor test (CPT)