hnp (version 1.2-7)

wolbachia: Trichogramma galloi parasitism data

Description

Viability of Trichogramma galloi, a parasitoid wasp, when infected with Wolbachia, a bacteria known to change its reproductive aspects.

Usage

data(wolbachia)

Format

A data frame with 106 observations on the following 3 variables.

y	numeric; number of eggs with an orifice	m
numeric; total number of parasitised eggs	treat	a factor with levels m+f+, m+f-, f+, m-f- and f-,
	where m stands for male, f stands for female, + means infected and - means non-infected;

Details

The bacteria Wolbachia is commonly found in various insect species and has the ability to change reproductive aspects of its host. When it infects the wasp Trichogramma galloi it is known to induce thelythokous parthenogenesis, i.e., only females are produced from unfertilized eggs. Souza (2011) conducted an experiment to assess the effects of Wolbachia on the viability of T. galloi eggs. Around 100 Diatraea saccharalis eggs (the host) were offered to infected (+) or non-infected (-) parasitoid couples or virgin females every day until the death of the female. The parasitised eggs, easily identifiable because they become dark, were then kept on moist filter paper for twenty days when counts were then made of the number of eggs that had an orifice, which meant that an adult parasitoid had emerged and thus the parasitoid was viable.

References

Souza, A. R. (2011) A interacao Wolbachia - Trichogramma galloi Zucchi, 1988 (Hymenoptera: Trichogrammatidae). Master's dissertation, ESALQ-USP

Examples

data(wolbachia)

# Binomial fit
model1 <- glm(cbind(y, m-y) ~ treat, family=binomial, data=wolbachia)
anova(model1, test="Chisq")

# Quasi-binomial fit
model2 <- glm(cbind(y, m-y) ~ treat, family=quasibinomial, data=wolbachia)
summary(model2)
anova(model2,test="F")

## half normal plots
par(mfrow=c(1,2),cex=1.2, cex.main=1.1)
hnp(model1, print=TRUE, pch=4, main="(a) Binomial",
    xlab="Half-normal scores", ylab="Deviance residuals")
hnp(model2, print=TRUE, pch=4, main="(b) Quasi-binomial",
    xlab="Half-normal scores", ylab='')

if (FALSE) {
# Beta-binomial fit                                                
### using package aods3
require(aods3)                                                     
                                                                   
model3 <- aodml(cbind(y, m-y) ~ treat, family='bb', data=wolbachia)
hnp(model3, print=TRUE, pch=4,                                     
    xlab="Half-normal scores", ylab='Deviance residuals')          

### using package VGAM
require(VGAM)

model3a <- vglm(cbind(y, m-y) ~ treat, family=betabinomial,
                data=wolbachia)
summary(model3a)
1/(1+exp(-coef(model3a)[2])) # phi estimate
hnp(model3a, data=wolbachia)
}

## for discussion on the analysis of this data set,
## see Demetrio et al. (2014)