scArches_reproducibility

Reproducing results from the scArches.

Getting Started

In the first step, you will need to download datasets, latent space representations, and pre-trained models to run the each notebook and reproduce the results.

All datasets are available in this drive directory. All latent space represenstations and pre-trained models are available in this drive directory. After you downloaded necessary datasets and latents, you can run the each notebook and reproduce the results.

Table of Notebooks

Data Analysis

Study	notebook path
Toy (Splatter)	notebooks/data_analysis/pancreas.ipynb
Pancreas	notebooks/data_analysis/pancreas.ipynb
Mouse Brain	notebooks/data_analysis/MouseBrain.ipynb
Tabula Senis Muris	notebooks/data_analysis/TabulaSenis.ipynb
Mouse Cell Atlas	notebooks/data_analysis/MouseCellAtlas.ipynb
Panorama	notebooks/data_analysis/panorama.ipynb
Covid-19	notebooks/data_analysis/Covid-19.ipynb

Notebooks (including paper plots)

Notebook	path
Toy dataset	notebooks/sample_runs/splatter.ipynb
Sample Pancreas training with classification	notebooks/sample_runs/train_Pancreas.ipynb
Method Comparison - Pancreas	notebooks/figures/methodComparison-pancreas.ipynb
Method Comparison - Mouse Brain	notebooks/figures/methodComparison-mousebrain.ipynb
Iterative Surgery - Pancreas (Alpha)	notebooks/IterativeSurgery_with_OutOfSample/OoS+IS_Pancreas_Alpha.ipynb
Iterative Surgery - Pancreas (Alpha + Gamma)	notebooks/IterativeSurgery_with_OutOfSample/OoS+IS_Pancreas_Alpha+Gamma.ipynb
Out of sample batch correction - Pancreas (Alpha + Gamma)	notebooks/OutOfSample/OutOfSample_Pancreas_Alpha.ipynb
Out of sample batch correction - Pancreas (Alpha + Gamma)	notebooks/OutOfSample/OutOfSample_Pancreas_Alpha+Gamma.ipynb
Integration and Classification - TabulaSenisMuris	notebooks/sample_runs/tabula_senis_muris.ipynb
Integration - Tabula Senis Muris with MouseCellAtlas (MCA)	notebooks/sample_runs/tabula_senis_mca.ipynb
Sample HCL MCA training	notebooks/sample_runs/HCL_MCA.ipynb
HCL MCA Integration Latent Analysis	notebooks/figures/HCL_analysis.ipynb
HCL MCA Integration UMAPs	notebooks/figures/HCL_UMAPs.ipynb
Sample COVID19 training	notebooks/figures/COVID19_training.ipynb
COVID19 Latent Analysis	notebooks/figures/covid_analyses.ipynb
COVID19 UMAPs	notebooks/figures/COVID19_UMAPs.ipynb

To run the notebooks and scripts you need following packages :

scnet, tensorflow, keras, scanpy, numpy, scikit-learn, matplotlib, scipy and splatter(R).

Benchmarks

Method	Notebook Path
mnnCorrect	notebooks/Benchmarks/mnnCorrect.ipynb
Conos	notebooks/Benchmarks/Conos.ipynb
Harmony	notebooks/Benchmarks/Harmony.ipynb
Liger	notebooks/Benchmarks/Liger.ipynb
Scanorama	notebooks/Benchmarks/Conos.ipynb
PCA	notebooks/Benchmarks/PCA.ipynb
Seurat	notebooks/Benchmarks/Seurat.ipynb

In the notebooks, the data is assumed to be in a folder named "data" in the same directory as the notebook.