Scroll to navigation

GMT-MUSIC-BMR-CALC-COVG(1p) User Contributed Perl Documentation GMT-MUSIC-BMR-CALC-COVG(1p)

gmt music bmr calc-covg

NAME

gmt music bmr calc-covg - Uses calcRoiCovg.c to count covered bases per-gene for each given tumor-normal pair of BAMs.

VERSION

This document describes gmt music bmr calc-covg version 0.04 (2018-07-05 at 09:17:13)

SYNOPSIS

gmt music bmr calc-covg --gene-covg-dir=? --roi-file=? --reference-sequence=? --bam-list=? --output-dir=? [--cmd-list-file=?] [--cmd-prefix=?] [--normal-min-depth=?] [--tumor-min-depth=?] [--min-mapq=?]

General usage:

 ... music bmr calc-covg \
    --bam-list input_dir/bam_list \
    --output-dir output_dir/ \
    --reference-sequence input_dir/all_sequences.fa \
    --roi-file input_dir/all_coding_exons.tsv

To create a list of commands that will allow the processing of each tumor-normal pair in parallel with an LSF job scheduler:

 ... music bmr calc-covg \
    --bam-list input_dir/bam_list \
    --output-dir output_dir/ \
    --reference-sequence input_dir/all_sequences.fa \
    --roi-file input_dir/all_coding_exons.tsv \
    --cmd_list_file parallelizable_commands \
    --cmd_prefix bsub

In the above case, the commands printed into the output file "parallelizable_commands" can be run in parallel. After they complete, rerun this script as printed directly below (--cmd_list_file and --cmd_prefix have been removed) to merge the parallelized calculations:

 ... music bmr calc-covg \
    --bam-list input_dir/bam_list \
    --output-dir output_dir/ \
    --reference-sequence input_dir/all_sequences.fa \
    --roi-file input_dir/all_coding_exons.tsv

REQUIRED ARGUMENTS

gene-covg-dir Text
Directory where per-sample gene coverage files are located
roi-file Text
Tab delimited list of ROIs [chr start stop gene_name] (See Description)
reference-sequence Text
Path to reference sequence in FASTA format
bam-list Text
Tab delimited list of BAM files [sample_name normal_bam tumor_bam] (See Description)
output-dir Text
Directory where output files and subdirectories will be written

OPTIONAL ARGUMENTS

cmd-list-file Text
A file to write calcRoiCovg commands to (See Description)
cmd-prefix Text
A command that submits a job to your cluster (See Description)
normal-min-depth Integer
The minimum read depth to consider a Normal BAM base as covered
tumor-min-depth Integer
The minimum read depth to consider a Tumor BAM base as covered
min-mapq Integer
The minimum mapping quality of reads to consider towards read depth counts

DESCRIPTION

This script counts bases with sufficient coverage in the ROIs of each gene in the given pairs of tumor-normal BAM files and categorizes them into - AT, CG (non-CpG), and CpG counts. It also adds up these base-counts across all ROIs of each gene for each sample, but covered bases that lie within overlapping ROIs are not counted more than once towards these total counts.

By default, this script runs a C-based tool named calcRoiCovg for each sample one after another, taking ~30 mins per sample to generate per-ROI covered base counts. If the results of calcRoiCovg for a sample already exists in the output subdirectory roi_covgs, re-calculation is skipped. This allows you to run your own calcRoiCovg jobs in parallel or on multiple machines (Keep reading).

Speed things up by running calcRoiCovg jobs in parallel: If a compute cluster or multiple machines are available, run this script twice as follows:

  • Define cmd-list-file and cmd-prefix to generate a file with commands that can be submitted to a cluster or run manually. These jobs will write per-ROI base counts in a subdirectory roi_covgs.
  • After all the parallelized calcRoiCovg jobs are completed, run this script again to add them up and generate the final per-gene base counts in a subdirectory gene_covgs. Remember to remove the cmd-list-file and cmd-prefix arguments or you will just be re-creating a list of commands.

ARGUMENTS

--roi-file
The regions of interest (ROIs) of each gene are typically regions targeted for sequencing or are merged exon loci (from multiple transcripts) of genes with 2-bp flanks (splice junctions). ROIs from the same chromosome must be listed adjacent to each other in this file. This allows the underlying C-based code to run much more efficiently and avoid re-counting bases seen in overlapping ROIs (for overall covered base counts). For per-gene base counts, an overlapping base will be counted each time it appears in an ROI of the same gene. To avoid this, be sure to merge together overlapping ROIs of the same gene. BEDtools' mergeBed can help if used per gene.
--reference-sequence
The reference sequence in FASTA format. If a reference sequence index is not found next to this file (a .fai file), it will be created.
--bam-list
Provide a file containing sample names and normal/tumor BAM locations for each. Use the tab- delimited format [sample_name normal_bam tumor_bam] per line. Additional columns like clinical data are allowed, but ignored. The sample_name must be the same as the tumor sample names used in the MAF file (16th column, with the header Tumor_Sample_Barcode).
--output-dir
Specify an output directory where the following will be created/written: roi_covgs: Subdirectory containing per-ROI covered base counts for each sample. gene_covgs: Subdirectory containing per-gene covered base counts for each sample. total_covgs: File containing the overall non-overlapping coverages per sample.
--cmd-list-file
Specify a file into which a list of calcRoiCovg jobs will be written to. These can be scheduled in parallel, and will write per-ROI covered base-counts into the output subdirectory roi_covgs. If cmd-list-file is left unspecified, this script runs calcRoiCovg per sample one after another, taking ~30 mins per sample, but it skips samples whose output is already in roi_covgs.
--cmd-prefix
Specify a job submission command that will be prefixed to each command in cmd-list-file. This makes batch submission easier. Just run the cmd-list-file file as a shell script to submit jobs. cmd-prefix is "bsub" if your cluster uses the LSF job scheduler, or "qsub" in Torque. Add arguments as necessary. For example, "bsub -M 4GB" sets a soft memory limit of 4GB.

LICENSE

Copyright (C) 2010-2011 Washington University in St. Louis.

It is released under the Lesser GNU Public License (LGPL) version 3. See the associated LICENSE file in this distribution.

AUTHORS

 Cyriac Kandoth, Ph.D.

SEE ALSO

genome-music-bmr(1), genome-music(1), genome(1)
2018-07-05 perl v5.26.2