CPP

Code for the Cell Point Process https://www.biorxiv.org/content/10.1101/2020.11.04.368001v1

Environment

CPP was implemented with pytorch 1.4.0 and numpy 1.18.1.

Usage

See Example_Notebook.ipynb for an example of how to simulate data, create a video representation, and estimate parameters to the model.

In this repo

Data (.mat files) used for this paper can be found at https://www.dropbox.com/sh/ctrb51chmkyfhlt/AABH1A1jBFrVSahljz7VSbWaa?dl=0

frontier

• run_frontier.py: Python code that randomly generates parameters, simulates data across different numbers of cells and peaks, and then estimates parameters from that data
• send_frontier.sh: A bash script for sending multiple run_frontier.py jobs to the clusters. We simulated with numpy random seeds 1-50.
• frontier-figs-Copy1.ipynb: Jupyter notebook for producing the figures seen in the paper
• learned- .npz: The output from 50 simulations

tapi

• run_tapi.py: Python code for loading data from tapi_dose.mat and estimating parameters for each well
• learned-103120.npz: The npz output of run_tapi.py
• figs_and_stats_Cop1.ipynb: Creates figures showing estimated paramters vs. tapi concentration
• lieklihood-test.ipynb: Creates graph of relative likelihood to a control model vs. tapi concentration
• spatial_lambda_ratio_plots: Creates graph of percent of conditional intensity vs tapi concentrations

drug

• run_drug_screen.py: Python code for loading data from FINAL_EQ.mat (Goglia, 2019) and estimating parameters for each well
• sample-matching-edited.xlsx: Excel table of metadata for each well
• param-control-fixed- - .npz: Output of run_drug_screen.py for different segments of FINAL_EQ.mat
• drug-figures-clean.ipynb: Jupyter notebook to produce figures showing 3d plot of different parameters for each class

gfp

• run_gfp.py: Python code for loading data from dataforcpp.mat and estimating parameters for each well
• learned-102920.npz: Output of run_gfp.py (all wells)
• gfp-figs-Copy1.ipynb: Jupyter notebook to view results. The GFP figure was created in biorender.

wound healing

• datadrugsnorm.mat: matlab workspace for pulse data from MDCK wound healing experiments
• run_wound_cppg_v6.py: Python code for loading data from datadrugsnorm.mat and estimating parameters for each well
• params-{drug}-exp-g-x10.npz: Output of run_wound_cppg_v6.py, where {drug} corresponds to pharmacological condition