1.generate-profiles

Generate Profiles

In this second module of the recipe, we perform a series of steps to convert single cell data into morphology profiles. The profiling_config.yaml controls all the pertinent details for this module.

Step 0 - Merge Single Cells

This step merges compartment data and metadata, filters cells and features, and outputs single cell morphology profiles.

Step 1 - Aggregate

This step combines single cell data by averaging morphology features to form "aggregate" profiles. We typically form aggregate profiles for both gene- and guide-level perturbations.

Step 2 - Normalize

This step performs a normalization scheme for the provided data level. The data levels include single cell, genes, and guides.

Step 3 - Feature Select

This step performs a series of feature selection steps to isolate the most pertinent morphology features in the data.