Today I was doing a quick ELISA and realized that the reader was old and produced the results in PDF only. This poses multiple problems as they need to be manually rewritten to a file and post-processed later. Additionally the optical density measurments in my software are automatically calculated into concentration based on some shady functions which I never understood.

I decided to change this and learn this and that about analysing ELISA results.

So the first thing that got me - OPTICAL DENSITY measurments were unaffected by the analysis type. So I had to use these raw values as my starting point. Extracting these was a matter of using Rpoppler package to extract pdf into txt and than using som gsub commands to get the exact values that I needed.

Secondly, the manual of my elisa recommended using the 4 parameter models and I had to look for a function which could facilitate this. `drc`

package has this covered.

I quickly wrote a function based on a post on research gate, and was able to get my results.

```
#' Calculate the concentration of a substrate from a known OPTICAL DENSITY
#'
#' This function will give a calculated concentration of a subsance in 96 well
#' plate given the optical density from a plate reader.
#'
#' @param OD Optical density of the STANDARD DILUTION [numeric vector]
#' @param conc KNOWN concentrations of the standard dilution [numeric vector]
#' @param OD_sample Optical densities of samples [named vector]
#' @return A data frame with OD, log fitted values and calculated concentrations
#' of the samples and a plot with a standard curve standard and blue samples
#' @author Wojciech Francuzik
#' @details This function is useful when working with old plate readers. It uses
#' the FOUR PARAMETER CURVE to fit the model to your standards.
#' @export elisa4ll
elisa4ll <- function(OD,conc,OD_sample){
require(drc)
logconc <-log10(conc)# log10 from conc
stdcrvdata <- data.frame(OD,conc,logconc)
fit<-drm(formula = OD ~ conc , data = stdcrvdata, fct = LL.4())
samples <- data.frame(OD=OD_sample)# data from mesurments
samples$loganswer<- fit$coefficients[4]*
(((-1* fit$coefficients[3]+samples$OD)/
(fit$coefficients[2]-samples$OD))^(1/ fit$coefficients[1]))
samples$conc <- 10^samples$loganswer
plot(fit)
lines(samples$loganswer,samples$OD, type="points", col="blue")
return(samples)
}
```

It is available in my rlabbook package (which is a slow WIP).

Basically if you provide this function with a set of OD reads from samples and standard curve along with the known concentrations of the standard dilution - you should get your concentrations calculated accordingly. It also plots a nice plot. There could be a tone more features but is it necessary?

Hope this helps someone!