Skip to main content
Fig. 4 | BMC Bioinformatics

Fig. 4

From: Guidelines for correlation coefficient threshold settings in metabolite correlation networks exemplified on a potato association panel

Fig. 4

Relative number of significantly changing bootstrapped NI and REC-CNs. The figure illustrates the relative number of significantly changing edge numbers in bootstrapped CNs. a NI-CNs relative change in edge number based on bootstrapping analysis at different r-values, at r = 1 to 0 at decrements of 0.1. b NI- relative change in edge number based on bootstrapping analysis bootstrapping analysis at different r-values, at r = 0.3 to 0.2 at decrements of 0.01. c REC-CNs relative change in edge number based on bootstrapping analysis at different r-values, at r = 1 to 0 at decrements of 0.1 d REC-CNs relative change in edge number based on bootstrapping analysis at different r-values, at r = 0.5 to 0.4 at decrements of 0.01. The confidence interval was estimated by the modified Cox test at different r-value to p-value combinations (see main text for details). Bootstrapping was run 100 times with 80% of the samples allowing replacement. Grey vertical lines separate correlation coefficients, red dashed vertical lines represent proposed correlation coefficient threshold interval. a Bootstrapped NI-CNs at r = 1 to 0 at decrements of 0.1; b Bootstrapped REC-CNs at r = 1 to 0 at decrements of 0.1; c Bootstrapped REC-CNs at r = 0.5 to 0.4 at decrements of 0.01

Back to article page