revealed evidence for differences between the common and minor constructs. The overall ANOVA F statistic was 3.95 (p=0.012), with significant differences between the control and common haplotypes (p=0.048) and between the common and minor haplotypes (p=0.023). Figure 3a shows the relative levels of expression for the constructs in P19 cells. In the SH-SY5Y human neuroblastoma cells, one-way ANOVA of all five constructs approached significance (F=2.03, p=0.078; Figure 3b). Although a step-wise difference in the long constructs was suggested by the bar graph, secondary ANOVA remained non-significant (F=2.10, p=0.106). However, after the SH-SY5Y cells were differentiated, a significant difference consistent with the P19 cell results emerged. Examining all five constructs, there was a significant difference associated with haplotype (F=7.68, p<0.001). Tukey’s post hoc provided evidence for differences between the long common and mixed haplotypes (p=0.023), the long common and minor haplotypes (p=0.065), the long mixed haplotype and each of the short constructs (p=0.005 and 0.006), the long minor haplotype and each of the short versions (p=0.014 in each), and suggestive evidence for differences between the common and minor haplotypes (p=0.065). The control vector was significantly different from the common (0.045), short common (0.009), and short minor (0.009) constructs. These data are
