2.7. Statistical Analysis

Experiments were performed in a randomized complete block design with 10 replicates (Petri dishes) per treatment. Five explants each for stigma/style and ovary and 25 explants for anther/filament were used per plate. The embryogenic response of explants, the effect of different culture media, and the effect of explant type were expressed as a percentage on a Petri-dish basis and recorded 6 months after explant incubation. Percentages of embryo germination were recorded 2 months from the incubation of somatic embryos on PGR-free medium. The percentage data were arcsine square-root transformed prior to analysis. The results were back-transformed and presented as mean ± standard error. To highlight statistically significant differences and possible interactions between explant, medium, and genotype, the multi-way analysis of variance (ANOVA) was performed (p ≤ 0.05). One-way ANOVA was performed when the interaction between the factors was not significant. The separation of the averages was performed by Tukey’s test (p ≤ 0.05).

For the molecular analysis, only bands showing consistent amplification within the range of 200 bp to 3.5 kb were considered. Polymorphic ISSR and RAPD markers were scored for the presence (1) or absence (0) of bands for all the somaclones analyzed. All reactions were repeated at least twice, and only well-resolved, distinct, polymorphic, and reproducible bands across all runs were considered for analysis. Bands with the same migration were considered homologous fragments, independently of their intensity. Smeared DNA fragments and weak bands, which could not be readily distinguished, were excluded. In the SSR analysis, peak intensity was considered and compared with the internal size standard to estimate allele size as described by De Michele et al. [58]. For morphological evaluation, the data collected were subjected to statistical analysis using standard deviations of the mean and thereafter to ANOVA (p ≤ 0.05). Prior to analysis, percentage data were arcsine square-root transformed. Statistical analysis was performed using SigmaStat 3.5 for Windows.