Remove lm from one group ggplot2

Question

I have a df:

MixA<-c(11.3296500, 3.7765500, 13.2179250, 1.8882750, 60.4248063, 906.3720938, 483.3984500, 1450.1953500, 0.4130875, 0.0590125)
TPM<-c(496.510998, 121.020332, 32.194033, 0.584421, 63.569152, 253.681165, 0.000000, 30487.460466, 0.000000, 0.000000)
test<-data.frame(MixA,TPM)

I want to log so I arbitarily replace the zeros and I want to plot these 'zero' points in a seperate colour so I group the data:

test[test == 0]<-0.01
test$group<-as.factor(ifelse(test$TPM==1.000000e-02,0,1))

When I plot with ggplot and it plots a lm for both groups but I want to remove the lm for the 'zero' group.

ggplot(test, aes(x=log10(test$MixA),y=log10(test$TPM),color=group))  + geom_point(size=3) +
  geom_smooth(method=lm) + scale_x_continuous(limits=c(-2.5, 3)) + scale_y_continuous(limits=c(-2.5, 5))  +
  theme(legend.position = "none")

I tried adding aes(group=1) to geom_smooth but this seems to shift the whole line??:

ggplot(test, aes(x=log10(test$MixA),y=log10(test$TPM),color=group))  + geom_point(size=3) +
  scale_x_continuous(limits=c(-2.5, 3)) + scale_y_continuous(limits=c(-2.5, 5))  +geom_smooth(aes(group=1), method="lm")+
  theme(legend.position = "none")

Would it work for you if you filter the data and then plot ? test %>% filter(group != 0) %>% ggplot() + aes(x=log10(MixA),y=log10(TPM),color=group) + geom_point(size=3) + geom_smooth(method=lm) + scale_x_continuous(limits=c(-2.5, 3)) + scale_y_continuous(limits=c(-2.5, 5)) + theme(legend.position = "none") — Ronak Shah
My understanding is that this doesn't plot the 'zero' group? I want to plot the 'zero' group but without the lm — zoe

J.C. J.C. · Accepted Answer · 2019-10-03T02:03:05

You can modify your geom_smooth() statement to reflect the groups for which you wish to plot a line.

ggplot(data = test, aes(x=log10(MixA),y=log10(TPM),color=group))  + geom_point(size=3) +
geom_smooth(data = subset(test, group == 1), aes(x= log10(MixA),y= log10(TPM),color=group), method=lm) + 
scale_x_continuous(limits=c(-2.5, 3)) + 
scale_y_continuous(limits=c(-2.5, 5))  +
theme(legend.position = "none")

This should give you the plot you want.

Best.

Remove lm from one group ggplot2

1 Answers