In koncina/elisar: Elisa 4PL regression

library(tidyverse)

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)
knitr::knit_hooks$set(pngquant = knitr::hook_pngquant)

elisar

The goal of elisar is to handle Tecan Sunrise excel exports which were modified to include the plate layout and sample identifications. It will perform a 4PL regression (using the drc library) and return a data frame containing the concentrations corresponding to the O.D. values.

Installation

You can install elisar from github with:

# install.packages("devtools")
devtools::install_github("koncina/elisar")

Example

Prepare the Excel files

1. Export Tecan sunrise results as MS Excel .xls files.
2. Add the layout (column names from 1 to 12 and row names from A to H) to fill in the IDs for each deposited sample.
   • Plate the layout plate on the same sheet, another sheet in the same file or even in another excel file.
   • The standard values are constructed with a common leading std_key id (defaults to STD but can be adjusted in the elisa_analyse() function) merged to a trailing concentration value. For example: 250, 500 and 1000 pg/ml standard points would be encoded as STD250, STD500 and STD1000 (see wells in rows A-G and columns 11 to 12 in the screenshot below).
3. Add an optional table to extend the identifications. The table should contain the mandatory column id listing all IDs reported in the layout. One can add as much columns as required to fully describe the data.

_Example of a modified Excel file: _

Import the file in R

Use the read_plate() to read in the O.D. values together with the IDs and extended informations as a data frame.\ A vector containing multiple files can be passed to read_plate(): These files can contain multiple data plates (O.D. values) but only a single layout and ID table should be present among all files.

library(tidyverse)
library(elisar)

elisar_example() %>%
  read_plate() %>%
  glimpse()

Perform the regression

elisar_example() %>%
  read_plate() %>%
  elisa_analyse() %>%
  glimpse()

The elisa_analyse() function performs a 4 parameter logistic regression (using drc::drm()) and returns a data frame with the calculated concentration values.\ A warning is displayed when O.D. values are outside the standard curve range. These values will be tagged as FALSE in the in_range column.

The extract_standard() function extracts the standard points from the data frame (converting the dose values encoded in the id column to numbers)

elisar_example() %>%
  read_plate() %>%
  extract_standard(concentration, od) %>%
  arrange(concentration)

To render the regression curve, the output of extract_standard() can be used by ggplot() (with the elisar::stat_4pl() layer to draw the drc::drm() 4PL regression model).

elisar_example() %>%
  read_plate() %>%
  extract_standard() %>%
  ggplot(aes(x = concentration, y = od)) +
  scale_y_log10() +
  scale_x_log10() +
  annotation_logticks(sides = "lb") +
  geom_point(size = 3, alpha = 0.5) +
  stat_4pl(color = "red", size = 1) +
  labs(title = "Standard curve example",
       x = "Concentration in pg/ml",
       y = "O.D. value") +
  theme_classic()

koncina/elisar documentation built on May 20, 2019, 12:55 p.m.