Additional file 9.
Similarity between spatiotemporal patterns. Panels (A and B) show the results of a randomization test performed to assess the effect of the initial cultures on the degree of similarity between population distributions. For each habitat we calculated the ratio of the average difference in population distributions of habitats inoculated from the same cultures (<dsame>) relative to the average difference to all habitats inoculated from different cultures (<ddifferent>): drelative=<dsame>/<ddifferent>. The red arrows indicate <drelative>, obtained by averaging log[drelative ] over all habitats of a given device type. The blue distribution shows the values of <drelative> obtained using 10.000 randomizations, where each population distribution was assigned to a randomly chosen habitat. Note that values of drelative were log transformed before averaging, the figure shows the back-transformed values. (A) Devices of type-1. (B) Devices of type 2. Note how in all cases the <drelative> for the real dataset (in red) is much lower than the <drelative> obtained from the randomized dataset (in blue). *** indicates p < 0.001. (C) Comparison of the degree of similarity observed in type-1 and 2 devices combined to that observed in devices of type-5. For both groups the differences between population distributions in habitats inoculated from the same culture set (dsame) and the difference between population distributions in habitats inoculated from different culture sets (ddifferent) is shown. Values of dsame and ddifferent obtained for habitats inoculated from the same culture sets were averaged together. N.S. indicates p > 0.05 in a Wilcoxon rank sum test (comparison of ddifferent between type 1 and 5 devices) or Wilcoxon signed rank test (comparison between dsame and ddifferent for type 5 devices).
Format: PDF Size: 123KB Download file
This file can be viewed with: Adobe Acrobat Reader
van Vliet et al. BMC Microbiology 2014 14:116 doi:10.1186/1471-2180-14-116