Description

Pipeline for Dose-Response Curve Analysis.

Description

Provides a variety of tools for assessing dose response curves, with an emphasis on toxicity test data. The main feature of this package are modular functions which can be combined through the namesake pipeline, 'runtoxdrc', to automate the analysis for large and complex datasets. This includes optional data preprocessing steps, like outlier detection, solvent effects, blank correction, averaging technical replicates, and much more. Additionally, this pipeline is adaptable to any long form dataset, and does not require specific column or group naming to work.

README.md

cran.r-project.org

toxdrc

toxdrc provides a streamlined workflow for analyzing dose-response data.
The package includes a set of modular functions that collectively:

import and clean raw experimental data
run quality checks
apply optional user-defined transformations or normalizations
fit multiple dose–response models
automatically select the best-fitting model
estimate specified effect measures across experimental groups or conditions

This makes toxdrc well-suited for handling complex or multi-factor experiments where dose–response relationships must be modelled consistently.

Getting Started

You can install the development version of toxdrc from GitHub with:

# install.packages("pak")
pak::pak("jsalole/toxdrc")

library("toxdrc")

This will get the package into your R session.

Examples

This package comes built in with two example datasets. toxresult is a small dataset that looks at the results of a toxicity test within a single 24-well plate. cellglow is much larger with 1080 observations. These demonstrate when this package is useful.

length(unique(interaction(toxresult$TestID, toxresult$Dye)))
#> [1] 1

length(unique(interaction(cellglow$TestID, cellglow$Dye)))
#> [1] 45

With one set of data, existing options like the drc package can be quite manageable. These can also be assessed using toxdrc through the many modular functions that exist, like blankcorrect(), normalizeresponse(), modelcomp(), getECx(), among many others. This package is most useful is through the pipeline function, runtoxdrc(). This pipeline splits up the dataset into the subsets for analysis.

data_01 <- runtoxdrc(
  dataset = cellglow,
  Conc = Conc,
  Response = RFU,
  IDcols = c("Test_Number", "Dye", "Replicate", "Type"),
  quiet = TRUE,
  qc = toxdrc_qc(),
  normalization = toxdrc_normalization(
    blank.correction = TRUE,
    normalize.resp = TRUE
  ),
  toxicity = toxdrc_toxicity(),
  modelling = toxdrc_modelling(
    quiet = TRUE,
  ),
  output = toxdrc_output()
)

data_02 <- runtoxdrc(
  dataset = cellglow,
  Conc = Conc,
  Response = RFU,
  IDcols = c("Test_Number", "Dye", "Replicate", "Type"),
  quiet = TRUE,
  qc = toxdrc_qc(),
  normalization = toxdrc_normalization(
    blank.correction = TRUE,
    normalize.resp = TRUE
  ),
  toxicity = toxdrc_toxicity(),
  modelling = toxdrc_modelling(
    quiet = TRUE,
  ),
  output = toxdrc_output(
    condense = TRUE
  )
)

While the pipeline only requires a few functions, it can be tailored to meet the needs of most assays. The toxdrc_config() help file indicates how these arguments can be used.

The data_01 provides intermediate information on the pipeline for each subset, but data_02 will return a simple dataframe summarizung the results.

summary(data_01)
#>                     Length Class  Mode
#> 83167.aB.A.Spiked   11     -none- list
#> 83256.aB.A.Spiked   11     -none- list
#> 83344.aB.A.Spiked   11     -none- list
#> 83475.aB.A.Spiked   11     -none- list
#> 83476.aB.A.Spiked   11     -none- list
#> 83167.CFDA.A.Spiked 11     -none- list
#> 83256.CFDA.A.Spiked 11     -none- list
#> 83344.CFDA.A.Spiked 11     -none- list
#> 83475.CFDA.A.Spiked 11     -none- list
#> 83476.CFDA.A.Spiked 11     -none- list
#> 83167.NR.A.Spiked   11     -none- list
#> 83256.NR.A.Spiked   11     -none- list
#> 83344.NR.A.Spiked   11     -none- list
#> 83475.NR.A.Spiked   11     -none- list
#> 83476.NR.A.Spiked   11     -none- list
#> 83167.aB.B.Spiked   11     -none- list
#> 83256.aB.B.Spiked   11     -none- list
#> 83344.aB.B.Spiked   11     -none- list
#> 83475.aB.B.Spiked   11     -none- list
#> 83476.aB.B.Spiked   11     -none- list
#> 83167.CFDA.B.Spiked 11     -none- list
#> 83256.CFDA.B.Spiked 11     -none- list
#> 83344.CFDA.B.Spiked 11     -none- list
#> 83475.CFDA.B.Spiked 11     -none- list
#> 83476.CFDA.B.Spiked 11     -none- list
#> 83167.NR.B.Spiked   11     -none- list
#> 83256.NR.B.Spiked   11     -none- list
#> 83344.NR.B.Spiked   11     -none- list
#> 83475.NR.B.Spiked   11     -none- list
#> 83476.NR.B.Spiked   11     -none- list
#> 83167.aB.C.Spiked   11     -none- list
#> 83256.aB.C.Spiked   11     -none- list
#> 83344.aB.C.Spiked   11     -none- list
#> 83475.aB.C.Spiked   11     -none- list
#> 83476.aB.C.Spiked   11     -none- list
#> 83167.CFDA.C.Spiked 11     -none- list
#> 83256.CFDA.C.Spiked 11     -none- list
#> 83344.CFDA.C.Spiked 11     -none- list
#> 83475.CFDA.C.Spiked 11     -none- list
#> 83476.CFDA.C.Spiked 11     -none- list
#> 83167.NR.C.Spiked   11     -none- list
#> 83256.NR.C.Spiked   11     -none- list
#> 83344.NR.C.Spiked   11     -none- list
#> 83475.NR.C.Spiked   11     -none- list
#> 83476.NR.C.Spiked   11     -none- list

head(data_02)
#>                    ID Effect Measure   Estimate Std. Error      Lower     Upper
#> 1   83167.aB.A.Spiked           EC50  4.4205652  2.2945587  0.8473618 23.061455
#> 2   83256.aB.A.Spiked           EC50 16.2647147  2.4803215 10.0109977 26.425033
#> 3   83344.aB.A.Spiked           EC50  5.9924242  1.6964279  2.4340586 14.752787
#> 4   83475.aB.A.Spiked           EC50  0.8838429  0.5481938  0.1227810  6.362371
#> 5   83476.aB.A.Spiked           EC50 28.8840432  2.7741255 21.2771968 39.210426
#> 6 83167.CFDA.A.Spiked           EC50 41.4848324 10.7602489 18.1718124 94.706641
#>   best_model_name effect
#> 1            LN.4   TRUE
#> 2            W2.4   TRUE
#> 3            W1.4   TRUE
#> 4            LL.4   TRUE
#> 5            W2.4   TRUE
#> 6            LN.4   TRUE

This package is also capable of making multiple point-estimates for each subset:

data_03 <- runtoxdrc(
  dataset = cellglow,
  Conc = Conc,
  Response = RFU,
  IDcols = c("Test_Number", "Dye", "Replicate", "Type"),
  quiet = TRUE,
  normalization = toxdrc_normalization(
    blank.correction = TRUE,
    normalize.resp = TRUE
  ),
  modelling = toxdrc_modelling(
    EDx = c(0.2, 0.5, 0.7),
    quiet = TRUE
  ),
  output = toxdrc_output(condense = TRUE)
)

head(data_03)
#>                  ID Effect Measure  Estimate Std. Error      Lower    Upper
#> 1 83167.aB.A.Spiked           EC20  3.235809   2.003172  0.4511939 23.20612
#> 2 83167.aB.A.Spiked           EC50  4.420565   2.294559  0.8473618 23.06146
#> 3 83167.aB.A.Spiked           EC70  4.995914   2.404170  1.0801938 23.10618
#> 4 83256.aB.A.Spiked           EC20 15.189664   2.536946  8.9270818 25.84561
#> 5 83256.aB.A.Spiked           EC50 16.264715   2.480321 10.0109977 26.42503
#> 6 83256.aB.A.Spiked           EC70 16.728695   2.453028 10.4904421 26.67659
#>   best_model_name effect
#> 1            LN.4   TRUE
#> 2            LN.4   TRUE
#> 3            LN.4   TRUE
#> 4            W2.4   TRUE
#> 5            W2.4   TRUE
#> 6            W2.4   TRUE

Authours

Jack Salole - Intial Work - jsalole

Acknowledgments

This package is an extenision for the drc package.
This package uses code suggested by Nel on stack overflow for mselect2, which updates mselect to work within a function.

r-toxdrc

toxdrc

Getting Started

Examples

Authours

Acknowledgments

Version

License

Status

Source

Homepage

Platforms (78)

toxdrc

Getting Started

Examples

Authours

Acknowledgments

Version

License

Status

Source

Homepage

Platforms78 (78)

Platforms (78)