to share all things tchic

1. workflows:
   This folder contains 3 sub-folders: 1. snakemake-workflow, 2. cell-selection, 3. snakemake-workflow.
   The first snakemake processes both chic and transcriptome data from fastq until count tables (or for transcriptome: .loom files with RNA velocity information). After this, the transcriptome data can be analysed locally using the notebooks in the analysis/ folder. The next two steps are specific for chic data. Step 2 uses information from the transcriptome to define similar (neighbour) cells and define bad cells in the chic data. Step 3 filters the chic data and removes bad cells, produces TSS and gene body count tables and creates pseudobulk bigwig files using your favourite clusters (e.g. cell types, time points, replicates).

2. pre-processing:
   Here you can find the jupyter notebook for analysis of the transcriptome. For this we use Scanpy and scVelo. Here you can also make a csv file with cell names x cell types for splitting of the chic bam file into pseudobulk files, and a csv file with umap coordinates for filtering out bad cells in the chic data.
   

3. data:
   Here you can find relevant data files, e.g. cell type cluster csv files to split bam files into pseudobulks.

4. figures:
   Here you can find the notebooks that were used to generate figures corresponding to the biorxiv paper.

For running the snakemake workflow on the cluster:
Please install a conda or python virtual environment and install all necessary dependencies in this environment. Note that this needs to be activated before running the snakemake. The following dependencies are necessary: SingleCellMultiOmics package, cutadapt, BWA or bowtie, samtools, STAR, velocyto.

For local analysis/running code in jupyter notebooks:
All notebooks are python-based and therefore require basic packages such as matplotlib, seaborn, pandas, and numpy.
All RNAseq analyses are based on Scanpy. For more information, see here. Tutorials, various ways of visualising the data and more can be found here.
RNA velocity is computed using scVelo, for more information see their documentation. Especially the tutorials can be useful for basic velocity analyis and various ways of visualising and modelling.
Trajectory inference using PAGA is based on this tutorial. We also use wishbone, for more information see here and here.