Table 1. Quantification of axon guidance defects in PRRG4 overexpression experiments.

Percent of segments demonstrating phenotypes
BP102 Phenotype	Wild type	<robo	robo	>robo	slit	Total
wild-type	99	1	0	0	0	100
sca::hRobo1	48	41	9	2	0	100
sca::hRobo1, PRRG4	7**	34	42***	16*	1	100

Stage 16 embryonic nerve cords were stained with BP102 and examined for the indicated genotypes. PRRG4 refers to an UAS-PRRG4 construct, hRobo1 to UAS-human-Robo1 and sca to the scabrous-GAL4 pan-neuronal driver. The embryonic CNS commissures are separated by two cell bodies visible with Nomarski optics per segment. The axon scaffold was scored for wild type morphology, thickened commissures or only one midline cell visible. “<robo”, stereotypical robo1 phenotypes in which the cell bodies separating the commissures are not visible or a minimal gap between the commissures “robo”, embryo segments in which the lateral constriction in width was greater than that of robo mutants “>robo” and segments in which the axons had collapsed onto the midline “slit”. See Fig 4 for examples. Total refers to the total number of segments scored for each genotype. Expression of PRRG4 by sca-GAL4 alone is predicted not to lead to any phenotypes, as multiple copies of these transgenes only produce a low frequency of phenotypes; enhancing the penetrance and expressivity of PRRG4 expression was the impetus for co-expression with hRobo1. The data for all three genotypes was analyzed with a Kruskal-Wallis one-way ANOVA and all three genotypes were found to be statistically different in all categories (p < 0.0004) except “slit”. The data for sca::hRobo1 and sca::hRobo1, PRRG4 were directly compared with a Mann-Whitney U test and the “wild type”, “robo” and “>robo” categories were statistically different (p = 0.001, 0.0001 and 0.014 respectively and indicated with asterisk in the table; * p < 0.05, ** p < 0.01, *** p < 0.001).

The underlying data are shown in S1 Data.