On the analysis of viability data: an example with Drosophila

Abstract

Larval competition experiments involving two wild type and eight mutant strains of Drosophila melanogaster have been carried out following the substitution procedure proposed by Mather and Caligari (1981). Our main goal has been to compare the competitive abilities of two phenotypically indistinguishable strains (wild and Oregon-R) by means of their responses with eight different mutants. Prior to the analyses of viability data, we have studied the normalizing effect of several transformations in order to determine which was best suited for the analyses. The differences found among the five transformations tested and the untransformed data were not very great. The folded power transformation (Mosteller and Tukey, 1977) was finally chosen.No constant pattern in the responses of the two wild type strains to the mutant competitors was detected. This leads us to conclude that the nature of the competition between the two wild type strains cannot be predicted from a knowledge of their competition with other strains.

Non‐additive effects between genotypes: Implications for competitive fitness assays

“…Our goal was not to explore differences among alternative reference strains with respect to variation in the strengths of effects and the fraction of reference strains that show genotype‐by‐genotype interactions for relative fitness. Nevertheless, similar results have been previously reported (Adell et al., 1990 ). The fitness component viability of two wild‐type strains cannot be predicted from their competition against reference strains.…”

“…Nevertheless, similar results have been previously reported (Adell et al, 1990). The fitness component viability of two wild-type strains cannot be predicted from their competition against reference strains.…”

On the use of tester stocks to predict the competitive ability of genotypes

“…The above inconsistencies apparently give support to the claims raised by Adell et al (1990). However, as /m does not statistically differ from 0 for the Control or Dahomey strains, and /3dspa does not differ significantly from 0 for the Dahomey strain, there is a problem about which variables should be included in the model.…”

“…Turning now to Adell et al's (1990) paper, the intergenotypic effects of the eight mutant strains they used on the two wild-type ones (wild and Oregon-R) were estimated by subtracting the monoculture and duoculture slopes (fm -fld) given that, in the substitution experimental design Ød measures both intra-and intergenotypic competitive effects on the performance of the indicator genotype (Mather & Caligari, 1983;De Miranda & Eggleston, 1987). However, the absence of detectable intragenotypic competition (/3,,, 0) was observed in the wild strain, and intragenotypic effects were the same as intergenotypic ones (/3d 0) in 19 per cent of the cases.…”

“…In a recent report Adell et al (1990) analysed the competitive abilities of two phenotypically indistinguishable wild-type strains of Drosophila melanogaster by competing them against eight different mutants. They used the substitution experimental design (Mather & Caligari, 1981) and found that there was inconsistency in the results of duocultures when any of the wild-type strains were compared to the mutants.…”