Uses nmr_pca_outliers_robust to perform the detection of outliers
Normalize the full spectra to the internal calibrant region, then exclude that region and finally perform PQN normalization.
pipe_load_samples(samples_dir, glob = "*0", output_dir = NULL)
pipe_add_metadata(nmr_dataset_rds, excel_file, output_dir)
pipe_interpolate_1D(nmr_dataset_rds, axis, output_dir)
pipe_exclude_regions(nmr_dataset_rds, exclude, output_dir)
pipe_outlier_detection(nmr_dataset_rds, output_dir)
pipe_filter_samples(nmr_dataset_rds, conditions, output_dir)
pipe_peakdet_align(
nmr_dataset_rds,
nDivRange_ppm = 0.1,
scales = seq(1, 16, 2),
baselineThresh = 0.01,
SNR.Th = -1,
maxShift_ppm = 0.0015,
acceptLostPeak = FALSE,
output_dir = NULL
)
pipe_peak_integration(
nmr_dataset_rds,
peak_det_align_dir,
peak_width_ppm,
output_dir
)
pipe_normalization(
nmr_dataset_rds,
internal_calibrant = NULL,
output_dir = NULL
)The directory where the samples are
A wildcard aka globbing pattern (e.g. *.csv) passed on to grep() to filter paths.
The output directory for this pipe element
The nmr_dataset.rds file name coming from previous nodes
An excel file name. See details for the requirements
The excel file can have one or more sheets. The excel sheets need to be as simple as possible: One header column on the first row and values below.
Each of the sheets contain metadata that has to be integrated. The merge (technically a left join) is done using the first column of each sheet as key.
In practical terms this means that the first sheet of the excel file MUST start with an "NMRExperiment" column, and as many additional columns to add (e.g. FluidXBarcode, SampleCollectionDate, TimePoint and SubjectID).
The second sheet can have as the first column any of the already added columns, for instance the "SubjectID", and any additional columns (e.g. Gender, Age).
The first column on each sheet, named the key column, MUST have unique values. For instance, a sheet starting with "SubjectID" MUST specify each subject ID only once (without repetitions).
The ppm axis range and optionally the ppm step. Set it to NULL for autodetection
A list with regions to be removed Typically:
exclude = list(water = c(4.7, 5.0))
A character vector with conditions to filter metadata.
The conditions parameter should be a character vector of
valid R logical conditions.
Some examples:
conditions <- 'Gender == "Female"'
conditions <- 'Cohort == "Chuv"'
conditions <- 'TimePoint %in% c("T0", "T31")'
conditions <- c(Cohort == "Chuv", 'TimePoint %in% c("T0", "T31")')
Only samples fullfilling all the given conditions are kept in further analysis.
Segment size, in ppms, to divide the spectra and search for peaks.
The parameter of peakDetectionCWT function of MassSpecWavelet package, look it up in the original function.
All peaks with intensities below the thresholds are excluded. Either:
A numeric vector of length the number of samples. Each number is a threshold for that sample
A single number. All samples use this number as baseline threshold.
NULL. If that's the case, a default function is used (nmr_baseline_threshold())
The parameter of peakDetectionCWT function of MassSpecWavelet package, look it up in the original function. If you set -1, the function will itself re-compute this value.
The maximum shift allowed, in ppm
This is an option for users, TRUE is the default value. If the users believe that all the peaks in the peak list are true positive, change it to FALSE.
Output directory from pipe_peakdet_align
A peak width in ppm
A ppm range where the internal calibrant is, or NULL.
This function saves the result to the output directory
This function saves the result to the output directory
This function saves the result to the output directory
This function saves the result to the output directory
This function saves the result to the output directory
Pipeline: Filter samples according to metadata conditions
Pipeline: Peak detection and Alignment
Pipeline: Peak integration
Pipe: Full spectra normalization
If there is no internal calibrant, only the PQN normalization is done.
Other import/export functions:
files_to_rDolphin(),
load_and_save_functions,
nmr_data(),
nmr_meta_export(),
nmr_read_bruker_fid(),
nmr_read_samples(),
nmr_zip_bruker_samples(),
save_files_to_rDolphin(),
save_profiling_output(),
to_ChemoSpec()
Other metadata functions:
nmr_meta_add(),
nmr_meta_export(),
nmr_meta_get_column(),
nmr_meta_get(),
nmr_meta_groups()
Other outlier detection functions:
nmr_pca_outliers_filter(),
nmr_pca_outliers_plot(),
nmr_pca_outliers_robust(),
nmr_pca_outliers()
Other peak detection functions:
nmr_baseline_threshold(),
nmr_detect_peaks_plot_overview(),
nmr_detect_peaks_plot(),
nmr_detect_peaks_tune_snr(),
nmr_detect_peaks(),
nmr_identify_regions_blood(),
nmr_identify_regions_cell(),
nmr_identify_regions_urine(),
nmr_integrate_regions()
Other alignment functions:
nmr_align_find_ref(),
nmr_align()
Other peak integration functions:
get_integration_with_metadata(),
nmr_identify_regions_blood(),
nmr_identify_regions_cell(),
nmr_identify_regions_urine(),
nmr_integrate_peak_positions(),
nmr_integrate_regions()
## Example of pipeline usage
## There are differet ways of load the dataset
dir_to_demo_dataset <- system.file("dataset-demo", package = "AlpsNMR")
# excel_file <- system.file("dataset-demo",
# "dummy_metadata.xlsx",
# package = "AlpsNMR")
# output_dir <- tempdir()
## Load samples with pipes
# pipe_load_samples(dir_to_demo_dataset,
# glob = "*.zip",
# output_dir = "../pipe_output")
## Another way to load it
# nmr_dataset <- nmr_read_samples_dir(dir_to_demo_dataset)
## Saving the dataset in a .rds file
# nmr_dataset_rds <- tempfile(fileext = ".rds")
# nmr_dataset_save(nmr_dataset, nmr_dataset_rds)
## Interpolation
# pipe_interpolate_1D(nmr_dataset_rds,
# axis = c(min = -0.5, max = 10, by = 2.3E-4),
# output_dir)
## Get the new path, based in output_dir
# nmr_dataset_rds <- paste(output_dir, "\", "nmr_dataset.rds", sep = "", collapse = NULL)
## Adding metadata to samples
# pipe_add_metadata(nmr_dataset_rds = nmr_dataset_rds, output_dir = output_dir,
# excel_file = excel_file)
## Filtering samples
# conditions <- 'SubjectID == "Ana"'
# pipe_filter_samples(nmr_dataset_rds, conditions, output_dir)
## Outlier detection
# pipe_outlier_detection(nmr_dataset_rds, output_dir)
## Exclude regions
# exclude_regions <- list(water = c(5.1, 4.5))
# pipe_exclude_regions(nmr_dataset_rds, exclude_regions, output_dir)
## peak aling
# pipe_peakdet_align(nmr_dataset_rds, output_dir = output_dir)
## peak integration
# pipe_peak_integration(nmr_dataset_rds,
# peak_det_align_dir = output_dir,
# peak_width_ppm = 0.006, output_dir)
## Normalization
# pipe_normalization(nmr_dataset_rds, output_dir = output_dir)