GCalignR provides simple functions to align peak lists obtained from
Gas Chromatography Flame Ionization Detectors (GC-FID) based on
retention times and plots to evaluate the quality of the alignment. The
package supports any other one-dimensional chromatography technique that
enables the user to create a peak list with at least one column
specifying retention times as illustrated below.
As with other software you need to get used to the input format which is shown in the illustration:
- Row 1: Sample names
- Row 2: Variable names
- Row 3-N: GC data
- Each block belongs to a sample as shown for sample A (green) and sample B (orange) above
The latest release v1.0.7 is on CRAN. Click here for an overview of past releases and a brief description of applied changes.
install.packages("GCalignR", dependencies = T)The developmental version, currently identical to the CRAN release, is available here:
if (!("devtools" %in% rownames(installed.packages()))) {
install.packages("devtools")
} else if (packageVersion("devtools") < 1.6) {
install.packages("devtools")
}
devtools::install_github("mottensmann/GCalignR", build_vignettes = TRUE)To get started read the vignettes:
browseVignettes("GCalignR")Basic usage of the main function to align peaks:
data: Path to a text file (see input format above), or list of data frames, each corresponding to a samplert_col_name: column name of retention time valuesmax_linear_shift: Here, no adjustment of systematic linear driftmax_diff_peak2mean: Here, sort all peaks strictly by retention timemin_diff_peak2peak: Here, try to merge peaks when rt differs by less than 0.1
library(GCalignR)
aligned <- align_chromatograms(data = peak_data[1:4], # list of data frame
rt_col_name = "time", # retention time
max_linear_shift = 0, #
max_diff_peak2mean = 0,
min_diff_peak2peak = 0.08)
#> Run GCalignR
#> Start: 2024-07-04 09:32:04
#>
#> Data for 4 samples loaded.
#> No reference was specified. Hence, a reference will be selected automatically ...
#>
#> 'C2' was selected on the basis of highest average similarity to all samples (score = 0.06).
#> Start correcting linear shifts with "C2" as a reference ...
#>
#> Start aligning peaks ... this might take a while!
#>
#> Merge redundant rows ...
#>
#> Alignment completed!
#> Time: 2024-07-04 09:32:07The parameter values above differ from the defaults shown in the paper
and the package vignette. In a nutshell, we now suggest in most cases to
set max_diff_peak2mean = 0. This way peaks are first simply sorted
based on the given retention time value and then purely
min_diff_peak2peak specifies which peaks will be evaluated for a
merge. Additionally, this enables the possibility for a considerable
boost in computation speed of the first alignment steps (available since
version 1.0.5, currently only on GitHub!)
If you encounter bugs or if you have any suggestions for improvement (for instance on how to speed up the algorithm!), just contact meinolf.ottensmann[at]web.de
Also I´m happy to provide help if you can´t get it to work. Usually it is easy to solve small problems. However, in order to simplify this process please send a short description of the problem along with the code you have been using as a script file (.R) together with a minimal example input file (.txt).
