SplitTSV

Here we discuss the use of the SplitTSV command, which separates the fragment files by cell type according to a reference file with mapped barcode - cell-type associations.

Usage

SplitTSV \
    '$sample_directory' \
    '$output_directory' \
    '$barcodes'

Input Format

Required arguments:
sample_directory # Path prefix that stores all datasets containing fragment files (e.g., ~/atac_seq_data/mouse)
output_directory # Path prefix to store all output split fragment files by dataset (e.g., ~/atac_seq_data/mouse_split)
barcodes         # Barcode file in .csv format. 
                   First column must include sample_id:barcode (e.g., GSM4594382:AAACGAAAGAACCCGA-1)
                   Seventh column must include the associated cell_type name (e.g., Somatotrophs)

There is also a SplitTSV_trajectory version of the function, which was adapted to split the neonatal mm10 fragment files by associated pseudotime trajectory rather than cell type. This function requires the exact same inputs, but requires a different function because the barcode file format provides sample_id:barcode in the second column and trajectory_id in the fourth column.

Output format

Datasets will be output in the output_directory folder, with the following named format:

~/atac_seq_data/mouse_split/{dataset_id}/{dataset_id}_{cell_type}.tsv.gz

A sample sheet with all dataset ID's will also be provided in the output folder under the name sample_sheet.txt. This file must not be deleted as it is a required input for the following 'MergeCelltype' function.