Code repo for the manuscript: "Inferring and perturbing cell fate regulomes in human cerebral organoids"

The repo is structured as follows:

• integration/ contains scripts and functions used to integrate the RNA-seq and ATAC-seq data
  • get_bipartite_matches.py finds matching cells given a h5ad file with a common embedding (in our case CCA) between the datasets.
  • matching.py contains helper functions for the matching script
• trajectory/ contains scripts to infer the trajectory graph based on RNA velocity and cellrank
• pando/ contains scripts and functions used to infer the gene regulatory network
  • atac.R contains functions to manipulate and plot ATAC-seq data
  • pseudocells.R contains the functions to construct pseudocells as used in Kanton et al. 2019
  • fit_multiome_glm.R contains a script to do model fitting for GRN inference. This is now also implemented in the R package Pando
  • models.R contains helper functions for model fitting.
• crop_seq/ contains scripts and functions used for the analysis of the CROP-seq data
  • get_guide_umis.py is a script to extract guide UMIs with high read support given a cellranger output (molecule_info.h5) using a GMM-based approach
  • ko_inference.R performs estimation of perturbation probabilities using a linear model-based approach as described in the Perturb-seq paper. It's essentially a R implementation of the approach used in MIMOSCA.
  • enrichment.R implements statistical tests for detection of composition changes
  • plots.R has some functions to plot CROP/Perturb-seq data.
• utils/ contaions general utility functions
  • de.R contains functions to perform differential expression and differential accessibility analysis
  • utils.R has some other useful functions
• figures/ contains scripts to generate the plots from the figures