I am trying to draw three differents non-linear regression with ggplot2 (like I did with graphpad below (dotted line) (because graphpad can't compare non-linear regression between groups):
So far, I drew this graph:
With the following code:
gp <- ggplot(datapoidsmono, aes(x = time, y = weight)) +
stat_summary(aes(color = group), fun.data="mean_sdl", fun.args = list(mult=1), geom="errorbar", position = "identity", size=0.5, width=0.2) +
stat_summary(fun.y = "mean", geom = "point", size=3, aes(shape=group,color=group)) +
scale_x_discrete(name = "Days after injection") +
scale_y_continuous(name = "Weight (g)", limits=c(0, 4000), breaks = seq(0, 4000,500)) +
scale_color_manual(values=c("green", "blue", "red"), name="Treatment", labels=c("A","B","C")) +
scale_shape_manual(values=c(15,16,17), name ="Treatment", labels=c("A", "B", "C")) +
ggtitle("Weight variation over time") + theme(plot.title = element_text(hjust = 0.5)) +
theme(legend.position = "right") +
theme(legend.background = element_rect(size=0.5, linetype="solid", color ="black", fill="white")) +
theme(axis.line.x = element_line(size = 0.5, color = "black"),axis.text.x = element_text(color="black", size = 12),axis.line.y = element_line(size = 0.5, color = "black"),axis.text.y = element_text(color="black", size = 12),axis.title = element_text(size =15, face="bold"),plot.title = element_text(size =20, face = "bold"),panel.grid.major = element_line(color = "#F1F1F1"),panel.grid.minor = element_blank(), panel.background = element_blank())
I can't figure out how to draw a non-linear regression for each groups.
The following code did not return any drawn line (no error either):
ggplot(datapoidsmono, aes(time, weight, color = group)) +
geom_point() +
stat_smooth(method = "lm", se=FALSE)
Nor did this one (found here):
ggplot(datapoidsmono, aes(x = time, y = weight, colour=group)) +
stat_smooth(method = 'nls', formula = 'y~a*exp(b*x)') +
stat_smooth(color = 1, method = 'nls', formula = 'y~a*exp(b*x)') +
geom_point(aes(fill=group))
Any idea or clue would be useful to continue! Thanks
As suggested by @PoGibas, I added "group=group" in aes inside first line which worked pefectly to draw the lines!
I tried mulltiple solution to get the perfect fit:
gp + ggplot(aes(group=group))
stat_smooth(method = "lm", formula = y ~ x, size = 1, se = FALSE,colour = "black") +
stat_smooth(method = "lm", formula = y ~ x + I(x^2),size = 1, se = FALSE, colour = "blue") +
stat_smooth(method = "loess", formula = y ~ x, size = 1, se = FALSE, colour = "red") +
stat_smooth(method = "gam", formula = y ~ s(x), size = 1, se = FALSE, colour = "green") +
stat_smooth(method = "gam", formula = y ~ s(x, k = 3), size = 1, se = FALSE, colour = "violet") +
stat_smooth(method = "auto", se=F, colour = "yellow")
But I figured that simply gp + stat_smooth()
did perfectly the job (the LOESS method is used).
Now I am trying to change aspect (to a dotted line) and color of the fit...
I tryed gp + stat_smooth(se=F, aes(fill = group))
but now I have another legend box and my lines are always with the same color...
I also tryed to add linetype=group
in the aes, but when I use scale_linetype_manual(values=c("dotted", "dotted", "dotted"))
, every line is dotted (included errorbar)
The complete code is:
ggplot(datapoidsmono, aes(x = time, y = weight, group=group, linetype=group)) +
stat_summary(aes(color = group), fun.data="mean_sdl", fun.args = list(mult=1), geom="errorbar", position = "identity", size=0.5, width=0.2) +
stat_summary(fun.y = "mean", geom = "point", size=3, aes(shape=group,color=group)) +
scale_x_discrete(name = "Days after injection") +
scale_y_continuous(name = "Weight (g)", limits=c(0, 4000), breaks = seq(0, 4000,500)) +
scale_color_manual(values=c("green", "blue", "red"), name="Treatment", labels=c("A","B","C")) +
scale_shape_manual(values=c(15,16,17), name ="Treatment", labels=c("A", "B", "C")) +
ggtitle("Weight variation over time") + theme(plot.title = element_text(hjust = 0.5)) +
theme(legend.position = "right") +
theme(legend.background = element_rect(size=0.5, linetype="solid", color ="black", fill="white")) +
theme(axis.line.x = element_line(size = 0.5, color = "black"),axis.text.x = element_text(color="black", size = 12),axis.line.y = element_line(size = 0.5, color = "black"),axis.text.y = element_text(color="black", size = 12),axis.title = element_text(size =15, face="bold"),plot.title = element_text(size =20, face = "bold"),panel.grid.major = element_line(color = "#F1F1F1"),panel.grid.minor = element_blank(), panel.background = element_blank()) +
stat_smooth(se=F, aes(fill = group)) +
scale_linetype_manual(values=c("dotted", "dotted", "dotted"))
Thanks to @PoGibas and this post, I added
group=group, color=group
in the aes of ggplot and it gave me a good result.
ggplot(datapoidsmono, aes(x = time, y = weight, group=group, color=group)) +
stat_summary(aes(color = group), fun.data="mean_sdl", fun.args = list(mult=1), geom="errorbar", position = "identity", size=0.5, width=0.2) +
stat_summary(fun.y = "mean", geom = "point", size=3, aes(shape=group,color=group)) +
scale_x_discrete(name = "Days after injection") +
scale_y_continuous(name = "Weight (g)", limits=c(0, 4000), breaks = seq(0, 4000,500)) +
scale_color_manual(values=c("green", "blue", "red"), name="Treatment", labels=c("A","B","C")) +
scale_shape_manual(values=c(15,16,17), name ="Treatment", labels=c("A", "B", "C")) +
ggtitle("Weight variation over time") + theme(plot.title = element_text(hjust = 0.5)) +
theme(legend.position = "right") +
theme(legend.background = element_rect(size=0.5, linetype="solid", color ="black", fill="white")) +
theme(axis.line.x = element_line(size = 0.5, color = "black"),axis.text.x = element_text(color="black", size = 12),axis.line.y = element_line(size = 0.5, color = "black"),axis.text.y = element_text(color="black", size = 12),axis.title = element_text(size =15, face="bold"),plot.title = element_text(size =20, face = "bold"),panel.grid.major = element_line(color = "#F1F1F1"),panel.grid.minor = element_blank(), panel.background = element_blank()) +
stat_smooth(se=F, linetype="dotted")
NB: the fit proposed by graphpad (see first graph) is more stat_smooth(method = "lm", formula = y ~ x + I(x^2),size = 1, se = FALSE)
than the one I finally chose (LOESS)