How to create a (total) Ghrelin calibration curve?

Version 8
by (unknown)
Version 23
by (unknown)

Deletions or items before changed

Additions or items after changed

1 == Computing Ghrelin calibration curve using R ==
2 +
3 +
===This article describes creating calibration curve, and measuring the total-ghrelin concentration ===
4
5 The calibration curve can be created using R, and libracy "drc". Optionally, one can use library "sfmisc" for formatting of the labels on plot axis.
6 +
7 +
We will assume a 4-parameter log-logistic model:
8 +
9 +
[[Image(LL4.png, 300px)]]
10 +
11 The R code is attached below.
12
13 The example assumes the data to be available in file "ghrelin_conc std_a std_b avg.csv"
14
15 The measured data:
16 -
{|
17 -
! Ghrelin (ng/ml) || Standard a || Standard b
18 -
|-
19 -
| 1000000 || -0.040596823 || -0.052699697
20 -
|-
21 -
| 100000 || 0.136105144 || 0.119766263
22 -
|-
23 -
| 10000 || 0.61356354 || 0.606906959
24 -
|-
25 -
| 1000 || 0.846543873 || 0.839887292
26 -
|-
27 -
| 100 || 0.887693646 || 0.88345764
28 -
|-
29 -
| 0 || 0.896770802 || 0.896165658
30 -
|}
31
32 +
|| '''Ghrelin (ng/ml)''' || '''Standard a''' || '''Standard b''' ||
33 +
|| 1000000 ||-0.040596823||-0.052699697 ||
34 +
|| 100000 ||0.136105144||0.119766263 ||
35 +
|| 10000 ||0.61356354||0.606906959 ||
36 +
|| 1000 ||0.846543873||0.839887292 ||
37 +
|| 100 ||0.887693646||0.88345764 ||
38 +
||0||0.896770802||0.896165658 ||
39
40 {{{
41
42 ##### Install libraries
43 install.packages("drc")
44 install.packages("sfsmisc")
45 require(drc)
46 library(sfsmisc)
47
48 ##### Read the data
49 hormone.data <- read.csv("ghrelin_conc std_a std_b avg.csv")
50 hormone.data <- hormone.data[,1:3]
51 colnames(hormone.data)[1:3] <- c("Concentration","Response_1", "Response_2")
52
53 ##### Reorganize the data
54 hormone.data <- reshape(hormone.data, varying=c("Response_1","Response_2"), direction="long", v.names=c("Response"))
55 hormone.data <- hormone.data[,c("Concentration", "Response")]
56
57 ##### Fitting the model (4-parameter log-logistic function)
58 hormone.data.model <- drm(Response ~ Concentration, data = hormone.data, fct = LL.4())
59 summary(hormone.data.model)
60
61 +
}}}
62 +
63 +
The resultant parameters of a log-logistic equation are:
64 +
65 +
{{{
66 +
Model fitted: Log-logistic (ED50 as parameter) (4 parms)
67 +
Parameter estimates:
68 +
Estimate Std. Error t-value p-value
69 +
b:(Intercept) 9.5057e-01 2.2294e-02 4.2638e+01 0
70 +
c:(Intercept) -7.6010e-02 6.9075e-03 -1.1004e+01 0
71 +
d:(Intercept) 8.9163e-01 3.3216e-03 2.6843e+02 0
72 +
e:(Intercept) 2.5221e+04 7.7727e+02 3.2448e+01 0
73 }}}
74
75 The calibration curve can be plotted using the commands below:
76
77 {{{
78 ##### Plotting a nice plot
79 par(pty="s", mar=c(5,5,1,1))
80 plot(hormone.data.model, type="confidence", cex.lab=2, axes=F, xlim=c(-10,10^6))
81 axis(side=1, at=hormone.data[1:6,1], labels=pretty10exp(hormone.data[1:6,1]), cex.axis=1.2)
82 axis(side=2, at=seq(0,1,0.2), labels=seq(0,1,0.2))
83 plot(hormone.data.model, type="all", add=T, pch=21, col="red", lwd=1, cex=2, bg="green")
84 }}}
85
86 +
[[Image(Ghrelin.png)]]
87
88 +
The parameters of the eqution can be plugged into the formula below (an inverse of the model), and used in Excel, or other spreadsheet program.
89
90 +
[[Image(LL4-inv.png, 230px)]]
91
92 +
However, the concentration can be also easily estimated in R using "ED" function of the "drc" library. The code below demonstrates the concentration estimated from the response of 0.1, assuming alpha=0.05. The code returns the estimation, the error, and the condfidence interval.
93 +
94 +
{{{
95 ##### Computing the concentration from the response, for instance for a response=0.1, and alpha=1-0.95
96 -
ED(hormone.data.model, respLev=0.1, interval="delta", type="absolute", level=0.95)
+
ED(hormone.data.model, respLev=0.1, interval="delta", type="absolute", level=0.95)
97 +
}}}