How to add regression line to faceted plots with coord_trans() axes?

[Thomas]

I would like to add regression lines to faceted plots that have a transformed axis. I am having trouble getting geom_abline() and coord_trans() to work together. I can imagine that this might not work, but I do not fully understand why it doesn't work. (example code adapted from stackoverflow). I am assuming the explanation for why this doesn't work and the solution for making it work are obvious, but I am not seeing them...

require(ggplot2)

require(plyr)

###generate fake data

df <- data.frame(x = c(1:98)/100)

df$y <- 2 + 3 * df$x + rnorm(98, sd = 40)

#############################################################

##create a function to extract linear model parameters from transformed data

#############################################################

lm_eqn = function(df){

  m = lm(y ~ asin(sqrtx, df);

  eq <- substitute(italic(y) == a + b %.% italic(x)*","~~italic(r)^2~"="~r2, 

                   list(a = format(coef(m)[1], digits = 2), 

                        b = format(coef(m)[2], digits = 2), 

                        r2 = format(summary(m)$r.squared, digits = 3)))

 c(a = format(coef(summary(m))[1,1], digits = 2), 

   b = format(coef(summary(m))[2,1], digits = 2),

  eq=as.character(as.expression(eq)))                 

}

df$group <- c(rep(1:2,49))

eq <- ddply(df,.(group),lm_eqn)

eq$a<-as.numeric(eq$a)

eq$b<-as.numeric(eq$b)

##################################################

#Untransformed axes displaying regression line from transformed data

##################################################

x11()

p<-ggplot(data = df, aes(x = x, y = y)) +

  #coord_trans(xtrans="asn")+

  geom_abline(aes(intercept=a, slope=b), data=eq)+

  geom_point()

p1 = p + geom_text(data=eq,aes(x = 0.25, y = 0,label=eq), parse = TRUE, inherit.aes=FALSE) + facet_grid(group~.)

p1 

#################################################

##Transformed axes but regression line won't display

###############################################

x11()

p<-ggplot(data = df, aes(x = x, y = y)) +

  coord_trans(xtrans="asn")+

  geom_abline(aes(intercept=a, slope=b), data=eq)+

  geom_point()

p1 = p + geom_text(data=eq,aes(x = 0.25, y = 0,label=eq), parse = TRUE, inherit.aes=FALSE) + facet_grid(group~.)

p1