Drug1 = a dataframe of drug concentration (V2) and measured response (response)
| response |
V2 |
| 102.5303783 |
1896 |
| 102.5050507 |
948.2 |
| 100.3063443 |
474.1 |
| 100.2961096 |
237.1 |
| 89.3452755 |
118.5 |
| 38.90199925 |
59.26 |
| 15.41198625 |
29.63 |
| 10.33549066 |
14.22 |
| 10.96631549 |
7.112 |
| 2.669522625 |
3.556 |
| -0.597990565 |
1.778 |
| -7.09484506 |
0.8297 |
Using the tidydrc package, I fit an LL.4 model:
drug1.LL4 <- tidydrc_model(drug1, V2, response, model = LL.4())
I plot the curve - looks perfect :
tidydrc_plot(drug1.LL4,ed50=FALSE) + scale_x_log10()
Each point is the mean. I want to add error bars (standard error of mean) based on my four replicates. I go back to a different dataset containing four replicates, and follow this tutorial I found to calculate standardard error . My result is a dataframe called tgc2:
| variable |
N |
value |
sd |
se |
ci |
| 0.8297 |
4 |
-7.09484506 |
2.350358363 |
1.175179181 |
3.739944644 |
| 1.778 |
4 |
-0.597990565 |
5.320693227 |
2.660346613 |
8.466410251 |
| 3.556 |
4 |
2.669522625 |
7.608666393 |
3.804333196 |
12.10708612 |
| 7.112 |
4 |
10.96631549 |
11.38786336 |
5.69393168 |
18.12063184 |
| 14.22 |
4 |
10.33549066 |
0.737991633 |
0.368995816 |
1.174309372 |
| 29.63 |
4 |
15.41198625 |
1.218442351 |
0.609221175 |
1.938813679 |
| 59.26 |
4 |
38.90199925 |
1.443234999 |
0.7216175 |
2.296508946 |
| 118.5 |
4 |
89.3452755 |
2.284661761 |
1.14233088 |
3.635406689 |
| 237.1 |
4 |
100.2961096 |
0.182323318 |
0.091161659 |
0.290117085 |
| 474.1 |
4 |
100.3063443 |
2.232914056 |
1.116457028 |
3.553064543 |
| 948.2 |
4 |
102.5050507 |
0.042179293 |
0.021089646 |
0.067116667 |
| 1896 |
4 |
102.5303783 |
0.065724222 |
0.032862111 |
0.104581904 |
Now I try to go back and add the error bars to the original data plot, by doing this :
tidydrc_plot(drug1.LL4,ed50=FALSE) + scale_x_log10() + geom_errorbar(aes(ymin=tgc2$value-tgc2$se, ymax=tgc2$value+tgc2$se, x=drug1.LL4[[1]][[1]]$d, y=drug1.LL4[[1]][[1]]$r))
However, what I get is this :
You can see that the y-position of the error bars is exactly correct. But the x-position is not.
How can I fix this?