Package 'staRdom'

Description

Import EEMs from generic csv files.

Usage

.eem_csv(file, col = "ex")

Arguments

Value

list with EEM data

Description

Add data of a PARAFAC model derived from multiway from EEMs

Usage

.trans_parafac(parafac, em, ex, samples, comp, const, norm_factors)

Arguments

Value

parafac model

Description

Samples from an eemlist that were not used in the modelling process are added as entries in the A-modes. Values are calculated using fixed B and C modes in the PARAFAC algorithm. B and C modes can be provided via a previously calculated model or as matrices manually.

Usage

A_missing(
  eem_list,
  pfmodel = NULL,
  cores = parallel::detectCores(logical = FALSE),
  adj_samples = TRUE,
  components = NULL,
  const = NULL,
  control = NULL,
  ...
)

Arguments

Details

This function can be used to calculate A modes (sample loadings) for samples that were previously excluded from the modelling process (e.g. outliers). The wavelengths in the samples and in the PARAFAC model must be identically. Surplus wavelengths in the samples can be adjusted automatically with the argument adj_samples = TRUE, surplus wavelengths in the PARAFAC model have to be mitigated by interpolating the missing wavelengths in the samples manually. If there is a mismatch in wavelengths, the function provides you with the necessary information. Another way to use it would be a recombination of components from different models and calculating the according sample loadings. Expecially the later application is experimental and results have to be seen critically! Nevertheless, I decided to supply this function to stimulate some experiments on that and would be interested in your findings and feedback.

Value

object of class parafac

Examples

data(eem_list)
data(pf_models)

A_missing(eem_list, pf4[[1]], cores = 2)

Description

Baseline correction for absorbance data

Usage

abs_blcor(abs_data, wlrange = c(680, 700))

Arguments

Value

data.frame

Examples

data(absorbance)
abs_data_cor <- abs_blcor(absorbance)

abs_data_cor1 <- abs_blcor(absorbance[1:2])

Fit absorbance data to exponential curve. drm is used for the fitting process.

Description

Fit absorbance data to exponential curve. drm is used for the fitting process.

Usage

abs_fit_slope(
  wl,
  abs,
  lim,
  l_ref = 350,
  control = drmc(errorm = FALSE, noMessage = TRUE),
  ...
)

Arguments

Value

numeric exponential slope coefficient

See Also

drm

Examples

data(absorbance)
abs_fit_slope(absorbance$wavelength,absorbance$sample1,lim=c(350,400),l_ref=350)

Description

Calculating slopes and slope ratios of a data frame of absorbance data.

Usage

abs_parms(
  abs_data,
  cuvle = NULL,
  unit = c("absorbance", "absorption"),
  add_as = NULL,
  limits = list(c(275, 295), c(350, 400), c(300, 700)),
  l_ref = list(275, 350, 300),
  S = TRUE,
  lref = FALSE,
  p = FALSE,
  model = FALSE,
  Sint = FALSE,
  interval = 21,
  r2threshold = 0.8,
  cores = parallel::detectCores(logical = FALSE),
  verbose = FALSE
)

Arguments

Details

The absorbance data is a data frame with the first column called "wavelength" containg the wavelength. Each other column contains the data from one sample. You can use absorbance_read to read in appropriate data.

The following spectral parameters are calculated:

• $S_{275-295}$ slope between 275 and 295 nm calculated with nonlinear regression

• $S_{350-400}$ slope between 350 and 400 nm calculated with nonlinear regression

• $S_{300-700}$ slope between 275 and 295 nm calculated with nonlinear regression

• SR slope ratio, calculated by $S_{275-295}/S_{350-400}$

• E2:E3 ratio $a_{250}/a_{365}$

• E4:E6 ratio $a_{465}/a_{665}$

• $a_{254}$ absorbance at 254 nm

• $a_{300}$ absorbance at 300 nm

Depending on available wavelength range, values might be NA. Additionally other wavelength limits can be defined. The slope ratio might fail in this case. For further details please refer to Helm et al. (2008).

Value

A data frame containing the adsorption slopes and slope ratios in column, one line for each sample.

References

Helms, J., Kieber, D., Mopper, K. 2008. Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter. Limnol. Oceanogr., 53(3), 955–969 https://aslopubs.onlinelibrary.wiley.com/doi/10.4319/lo.2008.53.3.0955

Examples

data(absorbance)

a1 <- abs_parms(absorbance, cuvle = 5, verbose = TRUE, cores = 2)
a2 <- abs_parms(absorbance, cuvle = 5,l_ref=list(NA,NA,NA), lref=TRUE, cores = 2) # fit lref as well

Description

Reading absorbance data from txt and csv files.

Usage

absorbance_read(
  absorbance_path,
  order = TRUE,
  recursive = TRUE,
  dec = NULL,
  sep = NULL,
  verbose = FALSE,
  cores = parallel::detectCores(logical = FALSE),
  ...
)

Arguments

Details

If absorbance_path is a directory, contained files that end on "csv" or "txt" are passed on to read.table. If the path is a file, this file is read. Tables can either contain data from one sample or from several samples in columns. The first column is considered the wavelength column. A multi-sample file must have sample names as column names. All tables are combined to one with one wavelength column and one column for each sample containing the absorbance data. Column and decimal separators are guessed from the supplied data. In some cases, this can lead to strange results. Plaese set 'sep' and 'dec' manually if you encounter any problems.

Value

A data frame containing absorbance data. An attribute "location" contains the filenames where each sample was taken from.

See Also

fread

Examples

absorbance_path <- system.file("extdata", "absorbance", package = "staRdom")
absorbance <- absorbance_read(absorbance_path, verbose = TRUE, cores = 2)

Description

Converting EEM data from class eem to data.frame.

Usage

## S3 method for class 'eem'
as.data.frame(x, row.names = NULL, optional = FALSE, gather = TRUE, ...)

Arguments

Value

A data frame containing the EEM data.

Examples

data(eem_list)

as.data.frame(eem_list[[1]])
as.data.frame(eem_list[[1]],gather=FALSE)

Description

According to dilution data absorbance is either multiplied by the according factor or the undiluted absorbance data is deleted. You can either specify the cor_data data table coming from eem_dilcorr or supply an eemlist, and the dilution data to created on the fly.

Usage

eem_absdil(
  abs_data,
  eem_list = NULL,
  dilution = NULL,
  cor_data = NULL,
  auto = TRUE,
  verbose = FALSE
)

Arguments

Value

data frame

Examples

# no appropriate exmaple data available yet

Description

Applying functions on EEMs

Usage

eem_apply(data, func, return = c("eemlist", "value"), ...)

Arguments

Details

The EEMs are passed on as first argument to func. Additionally, the vector of excitation wavelengths is passed on as ex and the emission wavelengths as em. Therefore, the supplied function has to allow these arguments. The easiest way would be ... (see example).

Value

eemlist or list

Examples

## define a function, that would divide a matrix by its maximum
# more general, if you want to return a valid eemlist (see below),
# a matrix of the same size has to be returned
# ... is used as a placeholder for any argument, important: em and
# ex wavelengths are passed on, so the function needs to take them as arguments,
# even if they are not used
norm_max <- function(x, ...){
  x/max(x)
}

# load example data
data("eem_list")

# normalise eems by the function defined above
norm_eems <- eem_apply(eem_list,norm_max,"eemlist")

# plot the results to see the difference
ggeem(norm_eems)

# define another function. what values were used to
# multiply the eems with?
norm_fac <- function(x, ...){
  1/max(x)
}

# return a list of factors used for normalisation
norm_factors <- eem_apply(eem_list,norm_fac,"value")

unlist(norm_factors)

# return list of em vectors.
# important: x needs to be in the first position, but
# is not used later!
extr_em <- function(x,em,...){
  em
}

em_vectors <- eem_apply(eem_list,extr_em,"value")

em_vectors

Description

The function tries to lead you to possible problems in your data.

Usage

eem_checkdata(
  eem_list,
  absorbance,
  metadata = NULL,
  metacolumns = NULL,
  correction = FALSE,
  error = TRUE
)

Arguments

Details

The returned list contains character vectors with sample names where possible problems were found: problem (logical, whether a severe problem was found), nas (sample names with NAs in EEM data), missing_correction (correction of EEM samples was not done or not done successfully),eem_no_abs (EEM samples with no absorbance data), abs_no_eem (samples with present absorbance but no EEM data), duplse (duplicate sample names in EEM data), duplsa (duplicate sample names in absorbance data), invalid_eem (invalid EEM sample name), invalid_abs (invalid absorbance sample name), range_mismatch (wavelength ranges of EEM and absorbance data are mismatching), metadupls (duplicate sample names in metadata), metamissing (EEM samples where metadata is missing), metaadd (samples in metadata without EEM data)

Value

writes out possible porblems to command line, additionally list with sample names where possible problems were found, see details.

Examples

folder <- system.file("extdata/EEMs", package = "staRdom") # load example data
eem_list <- eem_read(folder, recursive = TRUE, import_function = eem_csv)

abs_folder <- system.file("extdata/absorbance", package = "staRdom") # load example data
absorbance <- absorbance_read(abs_folder, cores = 2)

metatable <- system.file("extdata/metatable_dreem.csv",package = "staRdom")
meta <- read.table(metatable, header = TRUE, sep = ",", dec = ".", row.names = 1)

checked <- eem_checkdata(eem_list, absorbance, metadata = meta,
metacolumns = "dilution", error = FALSE)
# This example returns a message, that absorbance data for the
# blank samples are missing. As absorbance is supposed to be 0 over
# the whole spectrum when you measure blanks, there is no need
# to supply the data and do an inner-filter effect correction.

Description

The size of EEMs in an eemlist is checked and the sample names of samples with more data than the sample with the smallest range are returned.

Usage

eem_checksize(eem_list)

Arguments

Value

character vector

Examples

data(eem_list)
eem_checksize(eem_list)

Description

Return names of samples where certain corrections are missing.

Usage

eem_corrections(eem_list)

Arguments

Value

prints out sample names

Examples

data(eem_list)

eem_corrections(eem_list)

Description

This function can be used to import generic csv files containing EEM data using eem_read. Excitation wavelengths are assumed column-wise and emission wavelengths row-wise. If your data is arranged the other way round, please use eem_csv2

Usage

eem_csv(file)

Arguments

Value

list with EEM data

Examples

eems <- system.file("extdata/EEMs",package="staRdom")
eem_list <- eem_read(eems, recursive = TRUE, import_function = eem_csv)

eem_list

Description

This function can be used to import generic csv files containing EEM data using eem_read. Excitation wavelengths are assumed row-wise and emission wavelengths column-wise If your data is arranged the other way round, please use eem_csv

Usage

eem_csv2(file)

Arguments

Value

list with EEM data

Examples

## no example data provided with the package
## below is an example how this could like like
# eems <- "C:/some/path/to/eem.csv"
# eem_list <- eem_read(eems, recursive = TRUE, import_function = eem_csv2)

# eem_list

Description

Due to dilution absorbance spectra need to be multiplied by the dilution factor and names of EEM samples can be adjusted to be similar to their undiluted absorbance sample. The table contains information about these two steps. Undiluted samples are suggested by finding absorbance samples match the beginning of EEM sample name (see details).

Usage

eem_dilcorr(eem_list, abs_data, dilution, auto = FALSE, verbose = TRUE)

Arguments

Details

If you choose an automatic analysis EEMs are renamed if there is only one matching undiluted absorbance sample. Matching samples is done by comparing the beginning of the sample name (e.g. "sample3_1to10" fits "sample3").

Value

data frame

Examples

# no appropriate exmaple data available yet

Description

If samples were diluted before measuring, a dilution factor has to be added to the measured data. This function can do that by either multilpying each sample with the same value or using a data frame with different values for each sample.

Usage

eem_dilution(data, dilution = 1)

Arguments

Value

fluorescence data with class eemlist

Examples

data(eem_list)

eem_list2 <- eem_dilution(eem_list, dilution = 5)

dilutionT <- data.frame(dilution = rep(5, length(eem_list)))
row.names(dilutionT) <- eem_names(eem_list)
dilutionT

eem_list3 <- eem_dilution(eem_list, dilution = dilutionT)

Description

Check for duplicate sample names

Usage

eem_duplicates(data)

## Default S3 method:
eem_duplicates(data)

## S3 method for class 'eemlist'
eem_duplicates(data)

## S3 method for class 'data.frame'
eem_duplicates(data)

Arguments

Value

named character vector with duplicate sample names

Examples

### check

Description

In your default editor (e.g. RStudio), a Rmd file is opened. It consists of bloacks gathering the parameters and information needed and continues with a series of data corrections, peak picking and plots. Finally you get a report of your analysis, a table with the peaks and optional pngs of your fluorescence data. To continue working and keeping your settings, the file can be sa ved anywhere and reused anytime.

Usage

eem_easy()

Details

Function does not work well in Windows. You might try file.edit(system.file("EEM_simple_analysis.Rmd", package = "staRdom"))

Value

A pdf report, a peak picking table and optional plots.

Examples

## Not run: 
#
eem_easy()

# this function fails very often, so you might use that:
file.edit(system.file("EEM_simple_analysis.Rmd", package = "staRdom"))

## End(Not run)

Description

Correct names of EEM samples to match undiluted absorbance data.

Usage

eem_eemdil(
  eem_list,
  abs_data = NULL,
  dilution = NULL,
  cor_data = NULL,
  auto = TRUE,
  verbose = FALSE
)

Arguments

Value

eemlist

Examples

# no appropriate exmaple data available yet

Description

Outliers in all modes should be avoided. With this functions excitation or emission wavelengths as well as samples can be removed completely from your sample set.

Usage

eem_exclude(eem_list, exclude = list, verbose = FALSE)

Arguments

Details

The argument exclude is a named list of three vectors. The names must be "ex", "em" and "sample". Each element contains a vector of wavelengths or sample names that are to be excluded from the data set.

Value

object of class eemlist

Examples

data(eem_list)

exclude <- list("ex" = c(280,285,290,295),
"em" = c(),
"sample" = c("667sf", "494sf")
)

eem_list_ex <- eem_exclude(eem_list, exclude)

Description

Export all samples of an eem_list

Usage

eem_export(file, format = c("csv", "mat"), ...)

Arguments

Value

0 on successful export

Examples

# create temporary directory to write out
file <- paste0(tempdir(),"/eem_export/")
dir.create(file)

# run eem_export to write one csv file for each sample
eem_export(file, format = "csv", eem_list)

# show content of output directory
dir(file)

Description

Compared to the whole sample set, wavelengths missing in some samples are added and set NA or interpolated. This can be especially helpful, if you want to combine data measured with different wavelength intervals in a given range.

Usage

eem_extend2largest(eem_list, interpolation = FALSE, ...)

Arguments

Value

eemlist

Examples

library(dplyr)
data(eem_list)
eem_list <- eem_exclude(eem_list[1:5] %>%
`class<-`("eemlist"), exclude = list(em = c(318,322,326,550,438), ex = c(270,275))) %>%
eem_bind(eem_list[6:15] %>% `class<-`("eemlist"))
ggeem(eem_list)

eem_extend2largest(eem_list) %>%
  ggeem()

Description

Determines the the biggest range of EEM spectrum where data is available from each sample.

Usage

eem_getextreme(data)

Arguments

Value

list of numeric vector containing the biggest available range

Examples

data(eem_list)
eem_getextreme(eem_list)

eem_list <- eem_range(eem_list,ex = c(250,Inf),em = c(280,500))
eem_getextreme(eem_list)

Description

This function can be used to import txt files from Hitachi F-7000 containing EEM data using eem_read.

Usage

eem_hitachi(file)

Arguments

Value

list with EEM data

Examples

## no example data provided with the package
## below is an example how this could like like
# eems <- "C:/some/path/to/hitachi.TXT"
# eem_list <- eem_read(eems, recursive = TRUE, import_function = eem_hitachi)

# eem_list

Description

Calls eem_inner_filter_effect for each sample to use different cuvette lengths.

Usage

eem_ife_correction(
  data,
  abs_data,
  cuvl = NULL,
  unit = c("absorbance", "absorption")
)

Arguments

Value

fluorescence data of class eemlist

Examples

folder <- system.file("extdata/cary/scans_day_1", package = "eemR") # load example data
eem_list <- eem_read(folder, import_function = "cary")
data(absorbance)

eem_list <- eem_ife_correction(data = eem_list, abs_data = absorbance,
    cuvl = 5, unit = "absorbance")

Description

Reads Rdata and RDa files with one eemlist each. The eemlists are combined into one and returned.

Usage

eem_import_dir(dir, verbose = FALSE)

Arguments

Value

eemlist

Examples

## Not run: 
# due to package size issues no example data is provided for this function
# eem_import_dir("C:/some_folder/with_EEMS/only_Rdata_files")

## End(Not run)

Description

Missing EEM data can be interpolated. Usually it is the result of removing scatter or other parts where noise is presumed. Different interpolation algorithms can be used (see details).

Usage

eem_interp(
  data,
  cores = parallel::detectCores(logical = FALSE),
  type = TRUE,
  verbose = FALSE,
  nonneg = TRUE,
  extend = FALSE,
  ...
)

Arguments

Details

The types of interpolation are (0) setting all NAs to 0, (1) spline interpolation with mba.points, (2) excitation and emission wavelength-wise interpolation with pchip and subsequent mean, (3) excitation wavelength-wise interpolation with pchip and (4) linear interpolation in 2 dimensions with na.approx and again subsequent mean calculation. Calculating the mean is a way of ensuring NAs are also interpolated where missing boundary values would make that impossible. Using type = 1, extrapolation can be suppressed by adding the argument extend = FALSE.

Value

object of class eemlist with interpoleted spectra.

References

Elcoroaristizabal, S., Bro, R., García, J., Alonso, L. 2015. PARAFAC models of fluorescence data with scattering: A comparative study. Chemometrics and Intelligent Laboratory Systems, 142, 124-130 doi:10.1016/j.chemolab.2015.01.017

See Also

pchip, mba.points, na.approx

Examples

data(eem_list)
eem_list <- eem_list[1:6]
class(eem_list) <- "eemlist"

remove_scatter <- c(FALSE, TRUE, TRUE, TRUE)

remove_scatter_width = c(15, 10, 16, 12)

eem_list <- eem_rem_scat(eem_list, remove_scatter, remove_scatter_width)

eem_list <- eem_interp(eem_list, cores = 2)

ggeem(eem_list)

eem_list2 <- eem_setNA(eem_list, ex = 200:280, interpolate=FALSE)

ggeem(eem_list2)

eem_list3 <- eem_interp(eem_list2, type = 1, extend = TRUE, cores = 2)

ggeem(eem_list3)

eem_list3 <- eem_interp(eem_list2, type = 1, extend = FALSE, cores = 2)

ggeem(eem_list3)

Description

Check for NAs in EEM data

Usage

eem_is.na(eem_list)

Arguments

Value

named character vector with sample names where EEM data contains NAs

Examples

### check

Description

15 fluorescence samples from drEEM used for examples.

Usage

eem_list

Format

eemlist

Description

2 fluorescence samples from drEEM that were excluded as outliers from the PARAFAC model.

Usage

eem_list_outliers

Format

eemlist

Description

Load original data from the drEEM tutorial and return it as eemlist

Usage

eem_load_dreem()

Value

eemlist

Examples

# Reading MATLAB files from recent versions like the demo dataset from drEEM
# can cause problems if the R installation lacks UTF8 support in iconv.
# Therefore, we use try() in the example. If you encounter related problems,
# please refer to the help for R.matlab::readMat() for details.

eem_list <- try(eem_load_dreem(), silent = FALSE)
eem_list

Description

Multiply all EEMs with a matrix

Usage

eem_matmult(eem_list, matrix = NULL, value = 0)

Arguments

Details

All EEMs must be of the same size. If matrix is of type matrix, it is used right away to multiply the EEMs. It has to be of the same size as the EEMs. If matrix is a vector containing "l", values below 1st order Rayleigh scattering are set to 0. If matrix contains "u", values above 2nd order Raman scattering are set to 0. If you want to remove wavelength ranges, take into consideration to use eem_cut or eem_range.

Value

eemlist

Examples

data(eem_list)
eem <- eem_list[1:9]
class(eem) <- "eemlist"

ggeem(eem)

eem_list_cut <- eem_matmult(eem,matrix=c("l"), value= NA)
ggeem(eem_list_cut)

Description

You can use this table as an overview of your files and/or as a template for creating a metadata table.

Usage

eem_metatemplate(eem_list = NULL, absorbance = NULL)

Arguments

Value

data frame

Examples

folder <- system.file("extdata/EEMs", package = "staRdom") # load example data
eem_list <- eem_read(folder, recursive = TRUE, import_function = eem_csv)
data(absorbance)

eem_metatemplate(eem_list,absorbance)

Description

Sample names in eemlist can be altered.

Usage

eem_name_replace(eem_list, pattern, replacement)

Arguments

Details

str_replace_all from package stringr is used for the replacement. Please read the corresponding help for further options.

Value

An eemlist.

See Also

str_replace_all

Examples

eem_names(eem_list)

eem_list <- eem_name_replace(eem_list,"sample","Sample")
eem_names(eem_list)

Description

Plot fluorescence data from several samples split into several plots.

Usage

eem_overview_plot(data, spp = 8, ...)

Arguments

Value

list of ggplots

Examples

data(eem_list)
eem_overview_plot(eem_list, spp = 9)

# define number of rows and columns in plot
eem_overview_plot(eem_list, spp = c(3, 5))

Description

One or more PARAFAC models can be calculated depending on the number of components. The idea is to compare the different models to get the most suitable. B-mode is emmission wavelengths, C-mode is excitation wavelengths and, A-mode is the loadings of the samples. The calculation is done with parafac, please see details there.

Usage

eem_parafac(
  eem_list,
  comps,
  maxit = 2500,
  normalise = TRUE,
  const = c("nonneg", "nonneg", "nonneg"),
  nstart = 30,
  ctol = 10^-8,
  strictly_converging = FALSE,
  cores = parallel::detectCores(logical = FALSE),
  verbose = FALSE,
  output = "best",
  ...
)

Arguments

Details

PARAFAC models are created based on multiple random starts. In some cases, a model does not converge and the resulting model is then based on less than nstart converging models. In case you want to have nstart converging models, set strictly_converging TRUE. This calculates models stepwise until the desired number is reached but it takes more calculation time. Increasing the number of models from the beginning is much more time efficient.

Value

object of class parafac

See Also

parafac

Examples

data(eem_list)

dim_min <- 3 # minimum number of components
dim_max <- 7 # maximum number of components
nstart <- 25 # random starts for PARAFAC analysis, models built simulanuously, best selected
# cores <- parallel::detectCores(logical=FALSE) # use all cores but do not use all threads
cores <- 2 # package checks only run with 2 cores
maxit = 2500
ctol <- 10^-7 # tolerance for parafac

pfres_comps <- eem_parafac(eem_list, comps = seq(dim_min, dim_max),
    normalise = TRUE, maxit = maxit, nstart = nstart, ctol = ctol, cores = cores)

## with a defined number of converging models
#pfres_comps <- eem_parafac(eem_list, comps = seq(dim_min, dim_max),
#     normalise = TRUE, maxit = maxit, nstart = nstart, ctol = ctol,
#     output = "all", strictly_converging = TRUE, cores = cores, verbose = TRUE)

pfres_comps2 <- eem_parafac(eem_list, comps = seq(dim_min, dim_max),
    normalise = TRUE, maxit = maxit, nstart = nstart, ctol = ctol, cores = cores, output = "all")

Description

Calculate raman area of EEM samples

Usage

eem_raman_area(eem_list, blanks_only = TRUE, average = FALSE)

Arguments

Details

Code based on eem_raman_normalisation.

Value

data frame containing sample names, locations and raman areas

Examples

folder <- system.file("extdata/EEMs",package="staRdom")
eem_list <- eem_read(folder, recursive = TRUE, import_function = eem_csv)
blank <- eem_extract(eem_list,sample ="blank", keep = TRUE)

eem_raman_area(blank)

Description

Usually Raman normalisation is done with fluorescence data from a blank sample. Sometimes you already know a value for the Raman area. This function can do both.

Usage

eem_raman_normalisation2(data, blank = "blank")

Arguments

Details

Possible values for blank:

"blank": normalisation is done with a blank sample. Please refer to eem_raman_normalisation.

numeric: normalisation is done with one value for all samples.

data frame: normalisation is done with different values for different samples. Values are taken from a data.frame with sample names as rownames and one column containing the raman area values.

Value

fluorescence data of class eemlist

Examples

data(eem_list)
# correction by blank
eems_bl <- eem_raman_normalisation2(eem_list,blank="blank")

# correction by value
eems_num <- eem_raman_normalisation2(eem_list,blank=168)

Description

Cut EEM data matching a given wavelength range

Usage

eem_range(data, ex = c(0, Inf), em = c(0, Inf))

Arguments

Value

An eemlist of reduced spectra size.

Examples

data(eem_list)
eem_range(eem_list,ex = c(250,Inf),em = c(280,500))

Description

This function is deprecate, please use eem_read(..., import_function = eem_csv) or eem_read(..., import_function = eem_csv2) instead. EEM data is loaded from generic files. First column and first row contains wavelength values. The other values are to be plain numbers. fread is used to read the table. It offers a lot of helpful functions (e.g. skipping any number n of header lines by adding 'skip = n')

Usage

eem_read_csv(
  path,
  col = "ex",
  recursive = TRUE,
  is_blank_corrected = FALSE,
  is_scatter_corrected = FALSE,
  is_ife_corrected = FALSE,
  is_raman_normalized = FALSE,
  manufacturer = "unknown",
  ...
)

Arguments

Examples

eems <- system.file("extdata/EEMs",package="staRdom")
eem_list <- eem_read_csv(eems)

eem_list

Description

Remove wavelengths, that are missing in at least one sample form the whole set.

Usage

eem_red2smallest(data, return = "eems", verbose = FALSE)

Arguments

Details

This step is neccessary to perform a PARAFAC analysis which can only be calculated with spectra of similar range.

Value

eemlist with reduced spectral width

Examples

require(dplyr)

data(eem_list)

eem_list_red <- eem_red2smallest(eem_list)

# create an eemlist where data is missing
eem_list2 <- eem_exclude(eem_list,
    list("ex" = c(280,290,350),
         "em" = c(402,510),
         "sample" = c()))

# modify names of samples with missing data
eem_names(eem_list2) <- paste0("x",eem_names(eem_list2))

# combined the lists with and without missing data
eem_list3 <- eem_bind(eem_list,eem_list2)
#ggeem(eem_list3)

# reduce the data in the whole sampleset to the smallest wavelengths that are present in all samples
eem_list4 <- eem_red2smallest(eem_list3)
#ggeem(eem_list4)

Description

Wrapper function to remove several scatterings in one step using eem_remove_scattering.

Usage

eem_rem_scat(
  data,
  remove_scatter,
  remove_scatter_width = 10,
  interpolation = FALSE,
  cores = parallel::detectCores(logical = FALSE),
  verbose = FALSE
)

Arguments

Value

eemlist

Examples

data(eem_list)

remove_scatter <- c(TRUE, TRUE, TRUE, TRUE)

remove_scatter_width = c(15,10,16,12)

eems <- eem_rem_scat(eem_list,remove_scatter,remove_scatter_width)
ggeem(eems)

Description

Determine the range of fluorescence values in a set of samples

Usage

eem_scale_ext(data)

Arguments

Value

numeric vector

Examples

data(eem_list)
eem_scale_ext(eem_list)

Description

set parts of specific samples to NA and optionally interpolate these parts

Usage

eem_setNA(
  eem_list,
  sample = NULL,
  em = NULL,
  ex = NULL,
  interpolate = TRUE,
  ...
)

Arguments

Details

Samples and wavelengths are optional and if not set all of them are considered in setting data to NA. Wavelengths can be set as vectors containing more than the wavelengths present in the data. E.g. 230:250 removes all wavelengths between 230 and 250 if present. Data is best interpolated if it does not reach data boundaries. Please check the results otherwise as in some cases the interpolation might not produce meaningful data.

Value

eemlist

Examples

data(eem_list)
eem <- eem_list[1:9]
class(eem) <- "eemlist"

ggeem(eem)

eem_list2 <- eem_setNA(eem,ex=200:280,em=500:600, interpolate=FALSE)
ggeem(eem_list2)

Description

Smooth fluorescence data by calculating rolling mean along excitation wavelengths.

Usage

eem_smooth(data, n = 4, cores = parallel::detectCores(logical = FALSE))

Arguments

Value

eemlist with smoothed data

Examples

data(eem_list)

eem_list <- eem_smooth(eem_list, n = 4, cores = 2)

Description

Multiply EEMs with spectral correction vectors (Emission and Excitation)

Usage

eem_spectral_cor(eem_list, Excor, Emcor)

Arguments

Value

eemlist

Examples

eems <- system.file("extdata/EEMs",package="staRdom")
eem_list <- eem_read(eems, recursive = TRUE, import_function = eem_csv)

excorfile <- system.file("extdata/CorrectionFiles/xc06se06n.csv",package="staRdom")
Excor <- data.table::fread(excorfile)
emcorfile <- system.file("extdata/CorrectionFiles/mcorrs_4nm.csv",package="staRdom")
Emcor <- data.table::fread(emcorfile)

# adjust range of EEMs to cover correction vectors
eem_list <- eem_range(eem_list,ex = range(Excor[,1]), em = range(Emcor[,1]))

eem_list_sc <- eem_spectral_cor(eem_list,Excor,Emcor)

Description

Export samples in an EEM list to a single csv files

Usage

eem_write_csv(eem_list, output, ...)

Arguments

Value

returns the exported EEMs as a list of data.frames

Examples

data(eem_list)

output <- tempdir()
output
a <- eem_write_csv(eem_list, output)

Description

Data matrices from EEM are combined to an array that is needed for a PARAFAC analysis.

Usage

eem2array(eem_list)

Arguments

Value

object of class array

Examples

data(eem_list)

X <- eem2array(eem_list)

Description

Combining extracted components of PARAFAC models

Usage

eempf_bindxc(components)

Arguments

Value

parafac_components

Examples

data(pf_models)
pfmodel <- pf4[[1]]
comps <- eempf_excomp(pfmodel,c(1,3))
comps2 <- eempf_excomp(pfmodel,c(4,6))
comps3 <- eempf_bindxc(list(comps, comps2))

Description

Additionally a bar plot with the amounts of each component in each sample is produced.

Usage

eempf_comp_load_plot(pfmodel, ...)

Arguments

Value

ggplot

See Also

ggeem, eempf_load_plot

Examples

data(pf_models)

eempf_comp_load_plot(pf4[[1]])

Description

The components of a PARAFAC analysis are extracted as a data frame

Usage

eempf_comp_mat(pfmodel, gather = TRUE)

Arguments

Value

a list of class data frames

Examples

data(pf_models)

eempf_comp_mat(pf4[[1]])

Description

Extract names from PARAFAC model components

Usage

eempf_comp_names(pfmodel)

Arguments

Value

vector of names or list of vecters of names

Examples

data(pf_models)
eempf_comp_names(pf4)

eempf_comp_names(pf4) <- c("A","B","C","D","E","F","G")

value <- list(c("A1","B1","C1","D","E","F","G"),
c("A2","B2","C","D","E","F","G"),
c("A3","B3","C","D","E","F","G"),
c("A4","B4","C","D","E","F","G"),
c("A5","B5","C","D","E","F","G5")
)

eempf_comp_names(pf4) <- value
eempf_comp_names(pf4)

ggeem(pf4[[1]])

Description

Set names of PARAFAC components

Usage

eempf_comp_names(pfmodel) <- value

Arguments

Value

parafac model

Examples

data(pf_models)

eempf_comp_names(pf4) <- c("A","B","C","D","E","F","G")

Description

Three plots are returned:

1. plot of numer of components vs. model fit

2. plot of different components as colour maps

3. plot of different components as peak lines

The plots are intended to help with a suitable number of components.

Usage

eempf_compare(pfres, ...)

Arguments

Value

3 objects of class ggplot

See Also

eempf_fits, eempf_plot_comps

Examples

data(pf_models)

eempf_compare(pf4)

Description

Interactive 3D plots are created using plotly.

Usage

eempf_comps3D(pfmodel, which = NULL)

Arguments

Value

plotly plot

Examples

## Not run: 
data(pf_models)

eempf_comps3D(pf4[[1]])

## End(Not run)

Description

The convergence behaviour of all initialisations in a PARAFAC model is shown by printing the numbers

Usage

eempf_convergence(pfmodel, print = TRUE)

Arguments

Value

list with numbers of converging models, cflags and SSEs

Examples

data("pf_models")

pfmodel <- pf4[[1]]
conv_beh <- eempf_convergence(pfmodel)

Description

This is basically a wrapper for corcondia that deals with the normalisation of the original data., Other than corcondia, the default dicisor = "core".

Usage

eempf_corcondia(pfmodel, eem_list, divisor = "core")

Arguments

Value

numeric

Examples

## Not run: 
# due to data limitation in package, example does not work with that data!

# eempf_corcondia(pfmodel,eem_list)


## End(Not run)

Description

A pair plot showing correlations between samples is created.

Usage

eempf_corplot(
  pfmodel,
  normalisation = FALSE,
  lower = list(continuous = "smooth"),
  mapping = aes(alpha = 0.2),
  ...
)

Arguments

Value

object of class ggplot

See Also

ggpairs

Examples

data(pf_models)
eempf_corplot(pf4[[1]])

Description

Calculating correlations between the component loadings in all samples (C-Modes).

Usage

eempf_cortable(pfmodel, normalisation = FALSE, method = "pearson", ...)

Arguments

Value

matrix

Examples

data(pf_models)
eempf_cortable(pf4[[1]])

Description

Calculating EEMqual which is an indicator of a PARAFAC model's quality

Usage

eempf_eemqual(pfmodel, eem_list, splithalf = NULL, ...)

Arguments

Value

data frame containing fit, corcondia, product of best TCCs from splithalf analysis, eemqual and splithalf models

References

Rasmus Bro, Maider Vidal, EEMizer: Automated modeling of fluorescence EEM data, Chemometrics and Intelligent Laboratory Systems, Volume 106, Issue 1, 2011, Pages 86-92, ISSN 0169-7439

Examples

# data(eem_list)
# data(pf_models)

# pfmodel <- pf4[[1]]
# eempf_eemqual(eem_list,pfmodel) # insuficient example data to run!

Description

Extracting components of a PARAFAC model

Usage

eempf_excomp(pfmodel, comps)

Arguments

Value

list

Examples

data(pf_models)
pfmodel <- pf4[[1]]
comps <- eempf_excomp(pfmodel,c(1,3))

Description

Create one table containing the PARAFAC models factors and optionally exporting it to csv or txt

Usage

eempf_export(pfmodel, export = NULL, Fmax = TRUE, ...)

Arguments

Value

data frame

Examples

data(pf_models)

factor_table <- eempf_export(pf4[[1]])

Description

Fits vs. components of PARAFAC models are plotted

Usage

eempf_fits(pfres, ...)

Arguments

Value

object of class ggplot

Examples

data(pf_models)

eempf_fits(pf4)

Description

Calculate the leverage of each emission and excitation wavelength and each sample from a single PARAFAC model

Usage

eempf_leverage(pfmodel)

Arguments

Value

list of 3 named vectors (emission, excitation wavelengths and samples)

Examples

data(pf_models)

eempf_leverage(pf4[[1]])

Description

Combine leverages into one data frame and add optional labels.

Usage

eempf_leverage_data(cpl, qlabel = 0.1)

Arguments

Value

data frame

Examples

data(pf_models)

leverage <- eempf_leverage(pf4[[1]])
lev_data <- eempf_leverage_data(leverage)

Description

Plot is interactive where you can select values with your mouse. A list of vectors is returned to remove this outliers in a further step from your samples. The labels to be shown can be selected by adding the quatile of samples with highest leverages to be labeled.

Usage

eempf_leverage_ident(cpl, qlabel = 0.1)

Arguments

Value

list of three vectors containing the names of selected samples

See Also

eempf_leverage_plot

Examples

data(pf_models)

leverage <- eempf_leverage(pf4[[1]])
outliers <- eempf_leverage_ident(leverage)

Description

The labels to be shown can be selected by adding the quatile of samples with highest leverages to be labeled.

Usage

eempf_leverage_plot(cpl, qlabel = 0.1)

Arguments

Value

ggplot

See Also

eempf_leverage_ident

Examples

data(pf_models)

leverage <- eempf_leverage(pf4[[1]])
eempf_leverage_plot(leverage)

Description

Plot amount of each component in each sample as bar plot

Usage

eempf_load_plot(pfmodel)

Arguments

Value

ggplot

Examples

data(pf_models)

eempf_load_plot(pf4[[1]])

Description

Calculate the leverage of each emission and excitation wavelength and each sample from a list of PARAFAC models

Usage

eempf_mleverage(pfres_comps, ecdf = FALSE, stats = FALSE)

Arguments

Value

data frame containing leverages of wavelengths and samples for each model

Examples

data(pf_models)

eempf_mleverage(pf3)

Description

This function uploads a PARAFAC model to openfluor.org from within R. You need to have an account at openfluor.org and supply the email used for the account to the function. Your password is then asked in a secure way and only used within one execution of this function.

Usage

eempf_OF_upload(email, file)

Arguments

Value

HTTP status code from the upload POST

Examples

## due to the need of a valid account, this function cannot be
## tested with generic data.
## Please use your own account to do so.
## Not run: 
data(pf_models)

file <- file.path(tempdir(),"openfluor_example.txt")
eempf_openfluor(pf4[[1]],file)
eempf_OF_upload("[email protected]", file)

## End(Not run)

Description

openfluor.org offers the possibility to compare your results to others, that were uploaded to the database. This functions writes out a txt containing the header lines and your components. Please open the file in an editor and fill in further information that cannot be covered by this function.

Usage

eempf_openfluor(
  pfmodel,
  file,
  Fmax = TRUE,
  upload = FALSE,
  email = NULL,
  model_details = list()
)

Arguments

Value

txt file

Examples

data(pf_models)

  model_details <- list(name = "River", creator = "Helena Glory",
  constraints = "non-negative", validation = "split-half", unit= "RU")
  eempf_openfluor(pf4[[1]],file.path(tempdir(),"openfluor_example.txt"),
  upload = FALSE, model_details = model_details)

Description

The components can be plottet in two ways: either as a colour map or as two lines (emission, excitation wavelengths) intersecting at the component maximum. If the list of provided models is named, these names are shown in the plot. Otherwise, the models are automatically named by "model#".

Usage

eempf_plot_comps(
  pfres,
  type = 1,
  names = TRUE,
  contour = FALSE,
  colpal = "default",
  ...
)

Arguments

Value

object of class ggplot

Examples

data(pf_models)

eempf_plot_comps(pf4, type = 1)

# use a different colour scheme:
# eempf_plot_comps(pf4, type = 1, colpal = heat.colors(50))
eempf_plot_comps(pf4, type = 2)
eempf_plot_comps(list(pf4[[1]],pf4[[1]]), type=1)

Description

Plot results from an SSC check

Usage

eempf_plot_ssccheck(ssccheck)

Arguments

Value

ggplot element

Examples

data(pf_models)

ssccheck <- eempf_ssccheck(pfmodels = pf3[1:3], cores = 2)
eempf_plot_ssccheck(ssccheck)

Description

Reorder PARAFAC components

Usage

eempf_reorder(pfmodel, order, decreasing = FALSE)

Arguments

Value

parafac model

Examples

data(pf_models)
ggeem(pf4[[1]])

pf4r <- eempf_reorder(pf4[[1]], "ex")
ggeem(pf4r)

Description

Create a html report of a PARAFAC analysis

Usage

eempf_report(
  pfmodel,
  export,
  eem_list = NULL,
  absorbance = NULL,
  meta = NULL,
  metacolumns = NULL,
  splithalf = FALSE,
  shmodel = NULL,
  performance = FALSE,
  residuals = FALSE,
  spp = 5,
  cores = parallel::detectCores(logical = FALSE),
  ...
)

Arguments

Value

TRUE if report was created

Examples

folder <- system.file("extdata/EEMs", package = "staRdom") # load example data
eem_list <- eem_read(folder, recursive = TRUE, import_function = eem_csv)

abs_folder <- system.file("extdata/absorbance", package = "staRdom") # load example data
absorbance <- absorbance_read(abs_folder, cores = 2)

metatable <- system.file("extdata/metatable_dreem.csv",package = "staRdom")
meta <- read.table(metatable, header = TRUE, sep = ",", dec = ".", row.names = 1)

checked <- eem_checkdata(eem_list, absorbance, metadata = meta,
metacolumns = "dilution", error = FALSE)

eem_names(eem_list)
pfm <- A_missing(eem_list,pf4[[1]], cores = 2)
eempf_report(pfm, export = file.path(tempdir(),"pf_report.html"), eem_list = eem_list,
             absorbance = absorbance, meta = metatable, metacolumns = "dilution", cores = 2)

Description

B and C modes (emission and excitation wavelengths) are rescaled to RMS of value newscale. This is compensated in A mode (sample loadings).

Usage

eempf_rescaleBC(pfmodel, newscale = "Fmax")

Arguments

Value

object of class parafac

See Also

rescale

Examples

data(pf_models)

new_pf <- eempf_rescaleBC(pf4[[1]])

Description

Calculate residuals of EEM data according to a certain model

Usage

eempf_residuals(
  pfmodel,
  eem_list,
  select = NULL,
  cores = parallel::detectCores(logical = FALSE)/2
)

Arguments

Value

data frame with EEM residuals

Examples

data(eem_list)
data(pf_models)

residuals <- eempf_residuals(pf4[[1]], eem_list, cores = 2)

Description

The metrics calculated with this function are:

• RSS: residual sum of squares

• MAE: mean absolute error

• SAE: sum of absolute errors

• RSAE: sum of absolute error in relation to the sum of fluorescence and

• LEV: the leverage as described in eempf_leverage The example contains a way to plot these numbers.

Usage

eempf_residuals_metrics(residuals, leverage)

Arguments

Value

a list of data.frames containing residuals metrics for each sample, emission and excitation wavelength

Examples

data(eem_list)
data(pf_models)

residuals <- eempf_residuals(pf4[[1]], eem_list, cores = 2)
leverage <- eempf_leverage(pf4[[1]])

metrics <- eempf_residuals_metrics(residuals, leverage)

metrics$sample

## plot different residual metrics
require(dplyr)
require(tidyr)
require(ggplot2)

lapply(names(metrics), function(name){
  metrics[[name]] %>%
  mutate(mode = name, element = !!sym(name))
}) %>%
  bind_rows() %>%
  pivot_longer(cols = RSS:LEV, names_to = "metric", values_to = "value") %>%
  # uncomment the following line to select certain metrics
  # filter(metric %in% c("RSS","LEV")) %>%
  ggplot(aes(x = element, y = value, colour = metric))+
  geom_point()+
  facet_wrap(mode ~ ., ncol = 3, scales = "free")+
  theme(axis.text.x = element_text(angle = 90))+
  scale_y_continuous(trans="log")

Description

A raster of plots is created. Each column shows one sample. The top n rows show the n components from the model according their occurance in the certain samples. The second last row shows the residual, not covered by any component in the model and the last row shows the whole sample.

Usage

eempf_residuals_plot(
  pfmodel,
  eem_list,
  res_data = NULL,
  spp = 5,
  select = NULL,
  residuals_only = FALSE,
  cores = parallel::detectCores(logical = FALSE),
  contour = FALSE,
  colpal = "default"
)

Arguments

Details

eem_list may contain samples not used for modelling. Calculation is done by A_missing. This especially interesting if outliers are excluded prior modelling and should be evaluated again afterwards. Usually, residuals contain negative values, while these is the exception in samples and PARAFAC components. Therefore, we decided to use a similar colour palette as in the other plot functions but adding a purple tone for negative values.

Value

several ggplot objects

Examples

data(eem_list)
data(pf_models)

eem_list <- eem_extract(eem_list, 1:10)

eem_list <- eem_rem_scat(eem_list, rep(TRUE, 4), c(15,10,16,12))

eempf_residuals_plot(pf4[[1]], eem_list, cores = 2)

# use other colour schemes:
# eempf_residuals_plot(pf4[[1]], eem_list, colpal = c("blue",heat.colors(50)))
# plots <- eempf_residuals_plot(pf4[[1]], eem_list)
# lapply(plots, function(pl){
#   pl +
#     scale_fill_viridis_c() +
#     scale_colour_viridis_c()
# })

Description

The data variable pf_models can be supplied as list of PARAFAC models, output from a splithalf analysis or list of matrices Please see details of calculation in: U.J. Wünsch, R. Bro, C.A. Stedmon, P. Wenig, K.R. Murphy, Emerging patterns in the global distribution of dissolved matter fluorescence, Anal. Methods, 11 (2019), pp. 888-893

Usage

eempf_ssc(
  pfmodels,
  tcc = FALSE,
  m = FALSE,
  cores = parallel::detectCores(logical = FALSE)
)

Arguments

Value

(list of) tables containing SCCs between components

Examples

pf_models <- pf3[1:3]

sscs <- eempf_ssc(pf_models, cores = 2)
sscs

tcc <- eempf_ssc(pf_models, tcc = TRUE, cores = 2)
tcc
## mixed tcc (combine em and ex)
mtcc <- eempf_ssc(pf_models, tcc = TRUE, m = TRUE, cores = 2)
mtcc

## compare results from splithalf analysis
sh_sscs <- eempf_ssc(sh, cores = 2)

sh_sscs
## view diagonals only (components with similar numbers only)
lapply(sh_sscs, lapply, diag)

Description

Check SSCs between different models or initialisations of one model

Usage

eempf_ssccheck(
  pfmodels,
  best = length(pfmodels),
  tcc = FALSE,
  cores = parallel::detectCores(logical = FALSE)
)

Arguments

Value

data.frame containing SSCs

Examples

data(pf_models)

eempf_ssccheck(pf3[1:2], cores = 2)

# SSCs of split-half models, models need to be unlisted
data(sh)
eempf_ssccheck(unlist(sh, recursive = FALSE), cores = 2)

Description

Calculate the importance of each component.

Usage

eempf_varimp(
  pfmodel,
  eem_list,
  cores = parallel::detectCores(logical = FALSE),
  ...
)

Arguments

Details

The importance of each variable is calculated by means of creating a model without a specific component and calculating the difference between the original R-squared and the one with the left out component. The derived values state the loss in model fit if one component is not used in the modeling process. For the creation of the new models, the exact components of the original model are used.

Value

numeric vector, values are in the same order of the components in the supplied model.

Examples

data(pfmodel)
data(eem_list)

eempf_varimp(pf4[[1]], eem_list, cores = 2)

Description

Please refer to eem_biological_index, eem_coble_peaks, eem_fluorescence_index, eem_biological_index and abs_parms for details on the certain values

Usage

eempf4analysis(
  pfmodel,
  eem_list = NULL,
  absorbance = NULL,
  cuvl = NULL,
  n = 4,
  export = NULL,
  cores = parallel::detectCores(logical = FALSE),
  ...
)

Arguments

Value

data frame

Examples

data(eem_list)
data(pf_models)

results <- eempf4analysis(pfmodel = pf4[[1]],
                          eem_list = eem_list,
                          cuvl = 5, n = 4, cores = 2)

Description

Plots from EEM spectra of class ggplot. In case you work with a larger number of EEMs and want to show then in several plots, you can use eem_overview_plot.

Usage

ggeem(data, fill_max = FALSE, ...)

## Default S3 method:
ggeem(data, fill_max = FALSE, ...)

## S3 method for class 'eemlist'
ggeem(data, fill_max = FALSE, eemlist_order = TRUE, ...)

## S3 method for class 'eem'
ggeem(data, fill_max = FALSE, ...)

## S3 method for class 'parafac'
ggeem(data, fill_max = FALSE, ...)

## S3 method for class 'data.frame'
ggeem(
  data,
  fill_max = FALSE,
  colpal = "default",
  contour = FALSE,
  interpolate = FALSE,
  redneg = NULL,
  ...
)

Arguments

Details

The data can be of different sources: eem: a single EEM spectrum is plotted eemlist: all spectra of the samples are plotted, arranged in a grid data.frame: a data.frame containing EEM data. Can be created by e.g. as.data.frame.eem parafac: a PARAFAC model, the components are plotted then.

Using redneg you can give negative values a reddish colour. This can help identifying these parts in samples or components. Negative values are physically not possible and can only be the result of measuring errors, model deviations and problems with interpolated values.

Interpolation (interpolate = TRUE) leeds to smoother plots. The default is FALSE because it might cover small scale inconsistencies.

Contours (contour = TRUE)can be added to the EEM plots.

A colour palette can be specified using the argument colpal.

Plotting distinct samples can be done using eem_extract. Please see example.

Value

a ggplot object

Examples

## plotting two distinct samples
data(eem_list)
eem_names(eem_list)
eem <- eem_extract(eem_list,c("^d667sf$", "^d661sf$"),keep=TRUE)
ggeem(eem)

# the former redneg argument is deprecated, please see a similar looking example below!
#ggeem(eem, redneg = TRUE)
ggeem(eem, colpal = c(rainbow(75)[58],rainbow(75)[53:1]))

# use any custom colour palette
ggeem(eem, colpal = heat.colors(50))
# needs package matlab to be installed:
# ggeem(eem, colpal = matlab::jet.colors(50))
# or by adding ggplot2 colour and fill functions:
# ggeem(eem)+
#   scale_fill_viridis_c()+
#   scale_color_viridis_c()

ggeem(eem, interpolate = TRUE)
ggeem(eem, contour = TRUE)

Description

Full join of a list of data frames.

Usage

list_join(df_list, by)

Arguments

Value

The joint data frame.

See Also

full_join

Examples

a <- data.frame(what=letters[1:5],a=c(1:5))
b <- data.frame(what=letters[1:5],b=c(7:11))
c <- data.frame(what=letters[1:5],c=c(20:24))

df_list <- list(a,b,c)

list_join(df_list,by="what")

Description

Data for each wavelengths is returned. For each component the lines intersecting at the component maxima are returned.

Usage

maxlines(pfmodel)

Arguments

Value

data frame

Examples

data(pf_models)

ml <- maxlines(pf4[[1]])

Description

Normalise 3-dimensional array in first and second dimension

Usage

norm_array(eem_array)

Arguments

Value

array

Examples

data(eem_list)

a <- eem2array(eem_list)
an <- norm_array(a)

Description

Factors used for normalisation are saved separately in the PARAFAC models. With this function, the normalisation factors are combined with the A-modes of the model and removed as a separate vector. This means former normalisation is accounted for in the amount of each component in each sample. If no normalisation was done, the original model is returned without warning.

Usage

norm2A(pfmodel)

Arguments

Value

object of class parafac

Examples

data(pf_models)

pf4[[1]] <- norm2A(pf4[[1]])

Calculate a PARAFAC model similar to and using parafac.

Description

Please refer to parafac for input parameters and details. This wrapper function ensures 'nstart' converging models are calculated. On the contrary, parafac calculates 'nstart' models regardless if they are converging.

Usage

parafac_conv(
  X,
  nstart,
  verbose = FALSE,
  output = c("best", "all"),
  cl = NULL,
  ...
)

Arguments

Value

either a parafac model or a list of parafac models

See Also

parafac

Examples

data(eem_list)

dim_min <- 3 # minimum number of components
dim_max <- 4 # maximum number of components
nstart <- 25 # random starts for PARAFAC analysis, models built simulanuously, best selected
# cores <- parallel::detectCores(logical=FALSE) # use all cores but do not use all threads
cores <- 2 # package checks only run with 2 cores
maxit = 2500
ctol <- 10^-7 # tolerance for parafac

pfres_comps <- eem_parafac(eem_list, comps = seq(dim_min, dim_max),
    normalise = TRUE, strictly_converging = TRUE, maxit = maxit, nstart = nstart,
    ctol = ctol, cores = cores)

# keep all calculated models for diagnostics
pfres_comps_all <- eem_parafac(eem_list, comps = seq(dim_min, dim_max),
    normalise = TRUE, strictly_converging = TRUE, output = "all", maxit = maxit,
    nstart = nstart, ctol = ctol, cores = cores)

Description

PARAFAC model, see vignette, unconstrained

Usage

pf1

Format

list of parafacs

Description

PARAFAC model, see vignette, non-negative constraints

Usage

pf1n

Format

list of parafacs

Description

PARAFAC model, see vignette, non-negative constraints, normalised

Usage

pf2

Format

list of parafacs

Description

PARAFAC model, see vignette, non-negative constraints, normalised, outliers removed

Usage

pf3

Format

list of parafacs

Description

PARAFAC model, see vignette, non-negative constraints, normalised, outliers removed, high accuarcy

Usage

pf4

Format

list of parafacs

Description

result from PARAFAC split-half analysis, periodic data split

Usage

sh

Format

list of parafacs

Description

The samples are split into four subsamples: A,B,C,D. Subsamples are then combined and compared: AB vs. CD, AC vs. BD, AD vs. BC. The results show graphs from the components of each of the 6 models.

Usage

splithalf(
  eem_list,
  comps,
  splits = NA,
  rand = FALSE,
  normalise = TRUE,
  nstart = 20,
  cores = parallel::detectCores(logical = FALSE),
  maxit = 2500,
  ctol = 10^(-7),
  rescale = TRUE,
  strictly_converging = FALSE,
  verbose = FALSE,
  ...
)

Arguments

Details

Split data sets can be split suboptimal and cause low TCCs. Therefore, subsamples are recombined in 3 different ways and a TCC close to 1 in only one split combination per component is already a positive result. Check the split sets to check for sample independency.

Value

data frame containing components of the splithalf models

See Also

splithalf_plot, splithalf_tcc

Examples

data(eem_list)

splithalf <- splithalf(eem_list, comps = 6, verbose = TRUE, cores = 2)
splithalf_plot(splithalf)

# Similarity of splits using SSCs
sscs <- splithalf_tcc(splithalf)

Description

Graphs of all components of all models are plotted to be compared.

Usage

splithalf_plot(fits)

Arguments

Value

ggplot

See Also

splithalf

Examples

data(sh)

splithalf_plot(sh)
str(sh)

Description

Extracting a list of sample names in each subsample from a splithalf analysis

Usage

splithalf_splits(fits)

Arguments

Value

data frame containing TCC values

Examples

data(sh)
splithalf_splits(sh)

Description

Extracting TCC values from a splithalf analysis

Usage

splithalf_tcc(fits)

Arguments

Value

data frame containing TCC values

Examples

data(sh)

splithalf_tcc(sh)

Description

Please see details in: U.J. Wünsch, R. Bro, C.A. Stedmon, P. Wenig, K.R. Murphy, Emerging patterns in the global distribution of dissolved matter fluorescence, Anal. Methods, 11 (2019), pp. 888-893

Usage

ssc(mat1, mat2, tcc = FALSE)

Arguments

Value

table containing pairwise SCC of matrices columns

Examples

pf_models <- pf3
mat1 <- pf_models[[1]][[2]]
mat2 <- pf_models[[2]][[2]]

## calculate SSC
ssc(mat1,mat2)

## calculate TCC
ssc(mat1,mat2, tcc = TRUE)

Description

Calculate the combination of components giving the maximum of geometric mean of TCCs

Usage

ssc_max(mat)

Arguments

Value

vector with TCCs having the highest possible geometric mean

Examples

mat <- matrix(c(7,2,13,6,0,7,1,5,5), nrow = 3)
mat

sscs <- ssc_max(mat)
sscs

# order of components:
attr(sscs,"order")

Description

Componets must be passed as modes, see maxlines

Usage

tcc(maxl_table, na.action = "na.omit")

Arguments

Value

data.frame containing the TCCs

Examples

data(pf_models)

ml <- maxlines(pf4[[1]])

tcc(ml)

Description

When running a splithalf analysis similar components are not necessarily on the same position. This function looks for best fits with Tucker's Congruence Coefficients and returns a list of models with reordered components.

Usage

tcc_find_pairs(fits)

Arguments

Value

list of parafac models

See Also

splithalf

Examples

data(eem_list)

# function currently only used from within splithalf
splithalf(eem_list, 6, nstart = 2, cores = 2)