Post date: Jan 04, 2014 9:50:52 PM
The gemma analysis of survival with wild-caught individuals finished running (the mcmc results are in projects/lycaeides_hostplant/melGemma/mcmc_wild. I concatenated the predicted values across the ten chains and analyzed these in R. In particular, I found the wild-caught GLA had higher predicted survival on Ms and wild-caught SLA had higher survival on Ac (these are in the predicted direction and based on average predicted values across the 10 chains). The R code (in summarizeWildPrt.R) with results are below.
bvAc<-matrix(scan("cat_wildSurvAc.prdt.txt",n=6180),nrow=618,ncol=10,byrow=F)
bvMs<-matrix(scan("cat_wildSurvMs.prdt.txt",n=5660),nrow=566,ncol=10,byrow=F)
popAc<-c(rep(NA,520),rep("gla",40),rep("sla",58))
popMs<-c(rep(NA,468),rep("gla",40),rep("sla",58))
tapply(X=apply(bvAc,1,mean),INDEX=popAc,mean)
# gla sla
#0.05074267 0.84371371
out<-lm(apply(bvAc,1,mean) ~ as.factor(popAc))
anova(out)
# Df Sum Sq Mean Sq F value Pr(>F)
#as.factor(popAc) 1 14.8860 14.8860 6191.4 < 2.2e-16 ***
tapply(X=apply(bvMs,1,mean),INDEX=popMs,mean)
# gla sla
#0.6144346 0.5453283
#Residuals 96 0.2308 0.0024
out<-lm(apply(bvMs,1,mean) ~ as.factor(popMs))
anova(out)
# Df Sum Sq Mean Sq F value Pr(>F)
#as.factor(popMs) 1 0.11306 0.113057 7.6824 0.006698 **
#Residuals 96 1.41277 0.014716