1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
|
<!-- README.md is generated from README.rmd. Please edit that file -->
# chemCal - Calibration functions for analytical chemistry
<!-- badges: start -->
[](https://cran.r-project.org/package=chemCal)
[](https://app.travis-ci.com/jranke/chemCal)
[](https://codecov.io/github/jranke/chemCal)
<!-- badges: end -->
## Overview
chemCal is an R package providing some basic functions for conveniently
working with linear calibration curves with one explanatory variable.
## Installation
From within [R](https://www.r-project.org/), get the official chemCal
release using
``` r
install.packages("chemCal")
```
## Usage
chemCal works with univariate linear models of class `lm`. Working with
one of the datasets coming with chemCal, we can produce a calibration
plot using the `calplot` function:
### Plotting a calibration
``` r
library(chemCal)
m0 <- lm(y ~ x, data = massart97ex3)
calplot(m0)
```
<!-- -->
### LOD and LOQ
If you use unweighted regression, as in the above example, we can
calculate a Limit Of Detection (LOD) from the calibration data.
``` r
lod(m0)
#> $x
#> [1] 5.407085
#>
#> $y
#> [1] 13.63911
```
This is the minimum detectable value (German: Erfassungsgrenze),
i.e. the value where the probability that the signal is not detected
although the analyte is present is below a specified error tolerance
beta (default is 0.05 following the IUPAC recommendation).
You can also calculate the decision limit (German: Nachweisgrenze), i.e.
the value that is significantly different from the blank signal with an
error tolerance alpha (default is 0.05, again following IUPAC
recommendations) by setting beta to 0.5.
``` r
lod(m0, beta = 0.5)
#> $x
#> [1] 2.720388
#>
#> $y
#> [1] 8.314841
```
Furthermore, you can calculate the Limit Of Quantification (LOQ), being
defined as the value where the relative error of the quantification
given the calibration model reaches a prespecified value (default is
1/3).
``` r
loq(m0)
#> $x
#> [1] 9.627349
#>
#> $y
#> [1] 22.00246
```
### Confidence intervals for measured values
Finally, you can get a confidence interval for the values measured using
the calibration curve, i.e. for the inverse predictions using the
function `inverse.predict`.
``` r
inverse.predict(m0, 90)
#> $Prediction
#> [1] 43.93983
#>
#> $`Standard Error`
#> [1] 1.576985
#>
#> $Confidence
#> [1] 3.230307
#>
#> $`Confidence Limits`
#> [1] 40.70952 47.17014
```
If you have replicate measurements of the same sample, you can also give
a vector of numbers.
``` r
inverse.predict(m0, c(91, 89, 87, 93, 90))
#> $Prediction
#> [1] 43.93983
#>
#> $`Standard Error`
#> [1] 0.796884
#>
#> $Confidence
#> [1] 1.632343
#>
#> $`Confidence Limits`
#> [1] 42.30749 45.57217
```
## Reference
You can use the R help system to view documentation, or you can have a
look at the [online documentation](https://pkgdown.jrwb.de/chemCal/).
|
