Post-partition-reading Analysis

Typically the end-point analysis of dPCR data (i.e. the evaluation of negative vs positive partitions) is done by the software (eg., QX Manager, previously QuantaSoft) that is provided by the manufacturer of the dPCR equipment. Nevertheless, some advanced analyses, such as cluster gating (especially for higher plex data), can be performed with dedicated bioinformatics tools.

Below you will find some useful links to publications discussing dPCR data interpretation and new tools to analyze partition clusters:

  1. Polytect by Chen et al., 2025 is good at predicting and merging clusters when there are more clusters than channels and when one is doing higher orders of multiplexing! Run time per sample is mere seconds. Their (online) Shiny app can be accessed here: https://digpcr.shinyapps.io/Polytect/

  2. dPCP (R package) by De Falco et al., 2023 provides automated analysis of multiplex dPCR data generated by orthogonal and non-orthogonal assays using Bio-Rads’ QX and Thermos’ QuantStudio systems.

  3. ddPRclust (R package) & ddPCRvis (Online GUI / Shiny App) by Brink et al., 2018 were developed to analyze data from non-orthogonal (i.e. radial) multiplexed ddPCR (on Bio-Rads’ QX200 systems) reactions and their created cluster prediction.

  4. Finally, to get some general insight into the performance of different bioinformatics / software tools you could refer to a paper by Chen et al., 2024 wherein the authors benchmarked different dPCR data analysis tools.

For Bio-Rad’s QX systems, a .csv file containing amplitude data for every partition can be easily exported and analyzed further in software such as RStudio. An example of a complementary analysis (together with the code and example usage) to obtain: 1) 2D density plots with custom thresholds 2) comparison between experimentally obtained and “perfect” (here, perfectly normally distributed) simulated (from the experimental data) 2D clusters 3) and the theoretical predictions (using the experimental data) of the frequency distributions of droplets containing different target copy numbers (0, 1, 2 …) can be found here