In alexd106/ednar: Various Functions for Targeted qPCR eDNA Analysis

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#",
  fig.path = "man/figures/README-",
  out.width = "70%"
)

ednar

The goal of ednar is to provide a suite of easy to use functions to help with qPCR based targeted eDNA data analysis.

Installation

You can install the development version from GitHub with:

# install.packages("remotes")
remotes::install_github("alexd106/ednar")

Example usage

Plotting standard curves

To plot a basic qPCR standard curve for a specified target:

library(ednar)
calib_plot(calib_data, target = "706")

To include the the limit of detection (LOD) and optionally the limit of quantification (LOQ) (usually obtained using the calib_lod() function) on the plot use the lod = argument. LODs and LOQs can be supplied as either a data.frame object or as a vector.

# LOD as a vector
lod_vec <- 1.5
calib_plot(calib_data, target = "706", lod = lod_vec)

# LOD and LOQ as a vector
lod_vec <- c(1.5, 4.1)
calib_plot(calib_data, target = "706", lod = lod_vec)

# LOD and LOQ as a tibble
library(tibble)
lod_data <- tibble(Target = "706", lod = 1.5, loq = 4.3)
calib_plot(calib_data, target = "706", lod = lod_data)

The robust = TRUE argument will exclude standards with less than 50% detections when fitting the linear model

# LOD and LOQ as a tibble and robust = TRUE
library(tibble)
lod_data <- tibble(Target = "706", lod = 12.5, loq = 16.3)
calib_plot(calib_data, target = "706", lod = lod_data, robust = TRUE)

Summary statistics from standard curve data

To generate a table of summary statistics (slope and intercept estimates, test statistics etc) using standard curve data for each 'Target' we can use the calib_stats() function with the argument type = "effects".

sum_stats <- calib_stats(calib_data, type = "effects")
sum_stats

To obtain the R^2 estimates use type = "r2"

sum_stats <- calib_stats(calib_data, type = "r2")
sum_stats

Estimate LOD and LOQ from standard curve data

To obtain model based estimates of LOD and LOQ from the standard curve data from each 'Target' we can use the calib_lod() function. The calib_lod() function can automatically determine the 'best' model for LOQ by fitting linear, exponential decay, and up to 6^th order polynomial models and selecting the model with the lowest residual standard error. LOD estimates are obtained by fitting a series of dose response models using the drc package (see the drc::getMeanFunctions() function to see all possible models) and selecting the 'best' model based on log likelihood values, Akaike's information criterion and residual variance.

qpcr_lod <- calib_lod(data = calib_data, threshold = 0.35,
              lod.fit = "best", loq.fit = "best")

To display a summary of the LOD and LOQ estimates for each 'Target'

qpcr_lod$assaySum

To display a summary for each of the standards and Targets

qpcr_lod$standardsSum

With the calib_lod() function you can also fit specific models for both LOD and LOQ estimation. For example to estimate LOQ using an exponential decay model and LOD with a 4 parameter Weibull type II model.

qpcr_lod <- calib_lod(data = calib_data, threshold = 0.35,
              lod.fit = "W2.4", loq.fit = "decay")

alexd106/ednar documentation built on Sept. 1, 2021, 2:37 p.m.