The raw data matrix was square root transformed to downscale the contributions of quantitatively Z-IETD-FMK ic50 dominant FAs to the similarities (Clarke and Gorley 2006). A vector overlay was applied on the PCO plot to identify FA components responsible for differences between the three species based on Spearman’s correlation (r > 0.5). At each growth rate, the effect of N:P supply ratios on the content of each FA group (TFAs, SFAs, MUFAs, or PUFAs) and main individual PUFA (ALA, EPA, or DHA) was tested for each

algal species using one factorial ANOVA. The same analysis was done for the effect of growth rates on the content of each FA group and main individual PUFA under each N:P supply ratio. In the latter analysis, data for the contents of individual PUFAs were Ln (x) transformed. A post hoc test (Tukey’s HSD: Honestly Significant Difference) was applied only if there were significant effects. The magnitude of effect (ω2 = (effect sum of squares − effect degree of freedom × error mean square)/(total sum of squares + error mean square)) was calculated only for the significant factors. This

estimate can determine the variance in a response variable and relates this to the total variance in the response variable (Graham and Edwards 2001, Hughes and Stachowicz 2009). The relationship between the contents of FA groups Trametinib purchase (and main individual PUFAs) and cell quotas (QN and QP) was tested under the extremely N- and P-deficient conditions (N:P supply ratios = 10:1 and 63:1) using linear regression analyses. Data for QN and QP were published in our previous study (Bi et al. 2012). We compared FA profiles of the algal genus (Rhodomonas) and species (I. galbana and P. tricornutum) in this study with those of the same genus and species in the literature. All FA data were expressed as% of TFAs. Multidimensional scaling (MDS) and cluster analysis were conducted based on Bray–Curtis similarity resemblance medchemexpress matrix. The raw data matrix was

square root transformed. PCO, MDS, and cluster analysis were performed using the PERMANOVA+ add-on package to the PRIMER v6 software program (Clarke and Gorley 2006). ANOVA and linear regressions were conducted in Statistica 8 (StatSoft [Europe] GmbH, Hamburg, Germany). Significance level was set to P < 0.05 in all statistical tests. The equivalent spherical diameter (ESD) values for Rhodomonas sp. and P. tricornutum (Table 2) were obtained at the early stationary growth phase in batch cultures under N:P = 24:1, while that for I. galbana was not measured in this study. Cell densities and cellular C, N, and P contents showed both intra- and interspecific variations between the three algal species in semicontinuous cultures under different N:P supply ratios and growth rates. FA profiles varied between the three algal species. Tables S1–S3, in the Supporting Information, show the FA composition (as μg · mg C−1 and% of TFAs) for Rhodomonas sp., I. galbana, and P.