This pipeline is not under active development and NOT stable!
'linCNV' is a pipeline that can be called as a single executable program. This executable can be incorporated into the workflow management system of your preference.
The pipeline takes the output of 10x CellRanger DNA and uses the bam file to execute a different method of genome binning, based on bins with a fixed statistical weight.
The objective of the pipeline is to find high confidence copy number variant (CNV) calls in individual cells, and common CNVs in groups of cells, and to use that CNV information to assemble a cell lineage tree.
The pipeline calls a series of scripts in Perl, Bash and R, which sometimes call other programs on your system.
No installation is necessary beyond simply cloning or copying this folder of scripts to your computer.
Program dependencies can be examined using the executable (see help, below), which must be installed prior to use if not already present.
Use './linCNV' for command line help.
The pipeline supports parallel processing via options '--n-cpu' and '--ram-per-cpu'.
Option '--dry-run' allows a test of the command and option configuration prior to actual execution.
Commands 'options' and 'dependencies' provide further configuration assistance, while 'status' and 'rollback' provide support for monitoring and manipulating pipeline execution status.
First, the pipeline expects that you have already run the 10x Genomics Cell Ranger pipeline to complete at least read alignment, duplicate purging, and cell assignments. Point the pipeline to that Cell Ranger output using option 'cell-ranger-dir', where the code expects to find files named 'possorted_bam.bam' and 'per_cell_summary_metrics.csv'.
Additional required files define information about the genome in use. Option 'genome' might typically be 'hg38'. Option 'genome-dir' tells the pipeline where to find the genome reference files including '\<genome>.fa' and its index '\<genome>.fa.fai'.
The pipeline ignores low quality regions of the genome, which you must provide using option 'bad-regions-file', a BED format file that lists those offending regions. See https://pubmed.ncbi.nlm.nih.gov/31249361/ for details and where to obtain a suitable file for human and mouse.
Finally, 'gc-file' and 'mappability-file' must be set to two BED files that define genome bins that were assigned fraction GC and fraction mappability scores, respectively. These are used during bin data normalization. The size of bins in these files is not important, e.g. 1 kb bins.
Data files written by the pipeline are always placed into a folder you specify using option 'output-dir'. The file names are always prefixed with the value of option 'data-name'.
'output-dir' and 'data-name' are thus universally required options, and their values should be the same between sequential commands applied to the same data.
The general workflow occurs in the following conceptual steps:
1) The 'bin' command is run at the command line to create an initial set of genome bins.
2) The user individually examines and manually marks every cell for quality using the 'mark_cells' R Shiny web tool.
3) The 'analyze' command is run at the command line to refine the bins based on the user's cell marks and to complete data normalization and CNV calling.
4) The user examines the results using the 'heat_map' and other R Shiny web tools.
Instructions pending.