The current implementation computes summary statistics (e.g., mean or median) of marker expression across predicted cell areas. Often, the resulting values are log-transformed in downstream analyses. Reversing the order of operations from log(mean(x)) to mean(log(x)) can produce values that are more robust to the presence of outliers. However, there is no way to compute these values post-quantification.
Add a new command-line argument (e.g., --post-log) that guides the module to log transform the input intensities PRIOR to computing quantification statistics.
The current implementation computes summary statistics (e.g., mean or median) of marker expression across predicted cell areas. Often, the resulting values are log-transformed in downstream analyses. Reversing the order of operations from
log(mean(x))tomean(log(x))can produce values that are more robust to the presence of outliers. However, there is no way to compute these values post-quantification.Add a new command-line argument (e.g.,
--post-log) that guides the module to log transform the input intensities PRIOR to computing quantification statistics.