Sigmaplot

For the statistical analysis of data, SigmaPlot (version 14, SSI, San Jose, CA) was used. For the comparison of two samples from normally distributed populations with equal variances, Student’s t-test was performed, while in case of unequal variances a Kolmogorov-Smirnov test was applied. Multiple comparisons were performed with analysis of variance applying the Holm-Sidak test for post hoc pair-wise comparison of the data. In the case of unequal variances, the Dunnett T3 post hoc pair-wise comparison method was used. Differences were considered significant at p < 0.05.
All curve fitting was carried out by SigmaPlot (version 14, SSI, San Jose, CA) except for fitting of autocorrelation curves that was performed by using the QuickFit 3.0 software developed in the group B040 (Prof. Jörg Langowski) at the German Cancer Research Center (DKFZ).

Excel was used for basic data analysis and graphing while SigmaPlot 14 was used for conducting the analysis of variance. In SigmaPlot (Version 14.0), the one-way repeated-measures analysis of variance was used to compare the THC and CBD content of the bud material. SigmaPlot (Version 14.0) is preferred as it performs model adequacy checking and calculates the power of the test as well as generates the 95% confidence interval based on the standard error for the ratio of CBD to THC.

The presented data are expressed as mean ± standard deviation (SD). Sigma Plot (version14.0) and GraphPad Prism 7 software were used for statistical analysis and graphing, respectively. The data obtained from gene expression after the absolute quantification of mRNA by real-time polymerase chain reaction (qPCR) were analyzed to determine the normality and homogeneity of the variations using Shapiro–Wilk. When necessary, the data were transformed into log10 to achieve homogeneity and homogenized variances. Data obtained for cortisol and prostaglandin E2 were analyzed by one-way ANOVA, while fatty acids by two-way ANOVA, followed by a post-hoc Tukey HSD test for multiple comparisons. These last tests were also used for monoamine content analysis. All statistical analyses were performed using the SigmaPlot version 14 (SigmaPlot, Stata Statistical Software). Non-parametric statistics were also used when normality tests failed. Means and significance of test groups were compared to the controls. The significance indicator, or letter, above each graph represents a significant difference relative to control. A p < 0.05 was considered statistically significant.

For in vitro studies, n = 3 and
the statistical calculations within multiple groups
were done using ANOVA and Tukey’s, or ANOVA and Dunnett’s
multiple comparison test, as specified in the caption of each figure,
with 95% confidence interval using the GraphPad or SigmaPlot software.
Significant differences were reported as ***p <
0.001 unless reported otherwise in the figures.
For in vivo experiments, n = 6 animals per
experimental condition were used. Randomization was used to allocate
animals in different groups for the nude mice xenograft of CR-CSC
#21 using the EXCEL software method. For the tumor growth and weight
curve, the mean and SD of the tumor volume/animal weights at the end
point of each group were analyzed using two-way ANOVA and pairwise
multiple comparison procedures (Dunn’s method) post hoc nonparametric
test. The survival analysis of the animals was done using a Kaplan–Meier
survival plot in the SigmaPlot software. p value
was calculated with the log-rank test.

SSP density was defined as the mean number of patches m⁻². To obtain the patch mean density and associated standard error (SE) for each time period, each transect was divided into contiguous 20-m² bins (pseudoreplicates of 20 m of transect length by 1 m transect width). The bins were chosen in accordance with the ROV navigation data resolution as used in a recent megafauna study (Kuhnz et al. 2014 (link)). Abundances within bins were analyzed as replicate sample units to calculate mean densities. Temporal differences in density were assessed using a Kruskal–Wallis test (H, P, SigmaPlot version 12.5, Sokal and Rohlf 1995 ) as the data did not meet normality criteria. Correlations between SSP densities were also tested with living plate sponge densities during the same time period (rs, N, P, Spearman correlation, SigmaPlot version 12.5, Sokal and Rohlf 1995 ). The coefficient of dispersion (CD, variance/mean ratio) of individual patches was determined for each sampling period. By comparing the coefficient of dispersion of patches to a random distribution (Poisson distribution, CD = 1), the evenness of patch spatial distribution was assessed (Elliott 1971 ).