Normalization of MEMAs

Leveraging design principles to normalize high-throughput plate-based imaging data like MEMAs.

Our manuscript describing the method is available at Bioinformatics here

A docker image reproducing our analysis is available here

This github also contains our analysis files and R packages. The data is too large to host on github so one should alternatively see our docker image on dockerhub here or docker image/full files on zenodo here.

Analysis files

Most important analysis files are

1. process.ipynb (or mirrored process.R and process.Rmd)

This takes data from raw data to RR transformed data to RR and normalized data and saves it in various formats. Raw level 2 .tsv files are expected to be in the directory analysis/raw_data/MCF10A/

1. analysis_plots.ipynb (or mirrored analysis_plots.R and analysis_plots.Rmd)

This takes processed data and prodcues plots. This needs MCF10A_15_df.csv to be in the directory analysis/processed_data/csv/ but doesn’t require any of the lower-level data.

R packages

The underlying R package code for both the rr package and memanorm package may be found in the r_packages folder.