Background:
Elisa assays routinely provide results in the form of spreadsheets, one per assay plate,perhaps saved in folders by run date. Where the assay involves calibration against a standard using successive dilutions, the results for any specimen require the inspection of the calibration curves for their quality and their interpolation to obtain the concentration for the specimens being tested.

Material:
This problem arose in the context of a study of the effect of antiretrovial treatment of pregnant women in Kenya, where the levels of antibodies against rotavirus were of interest.
An Elisa assay was developed, where a standard, 10 specimens and a blank were placed in the columns of an Elisa plate at 8 successive dilutions. Data were obtained for a total of 788 specimens, which were run on 81 plates. Using an EMax endpoint microplate ELISA reader with SoftMax Pro for data acquisition and analysis produced a spreadsheet for each plate and a graph of titre vs. optical desnity (OD) in one worksheet and specimen details in another.

Methods:
We developed simple software tools to "scan" all the spreadsheets in the folder(s) of interest and assemble all the results from these sheets into a single data file. From these data, we used regression models to construct the calibration curves, assess these for their goodness of fit and interpolate the results where the linearity assumption is reasonable, as would usually be done by a human operator.

Work plan:
The purpose of this project is to improve the R and Stata code that we have developed, in order to provide a tool for laboratory scientists faced with such data. The end-product of the project should be a user-friendly tool, for example an on-line "app" using R-Shiny.

The project would be suited to a student with a laboratory science background, some statistical knowledge, and an interest in data handling using software.