Git repository of scripts useful for running bowtie alignments and aggregating read counts.
Script files are located in lib subdirectory. To run, execute with node ./lib/[script_name] or make script executable and execute with just ./lib/[script_name]
==================
run bowtie alignments against multiple sequence read files, aggregating read counts and summarizing alignments
================== required arguments:
- --fasta_file - path to fasta file to align against
- --fastq_directory - path to directory containing fastq files for alignment (if bowtie2 is used, these files may be .gz or .bz2 archives)
- --output_directory - path to directory to output resulting count files to
- --summary_file - path to save a summary file to, summarizing the results of bowtie alignments
================== optional arguments:
-
--bowtie_cmd - command to run bowtie, must be available in PATH or resolve to bowtie binary (default: "bowtie2")
-
--bowtie_args.[argument name (with - or -- if required)]=[argument value] - arguments to be passed to bowtie command. defaults:
-
-p=[parallel cores to use (all cores on your system)]
-
-x=[path to index files] *-U=[path to fastq file]
-
refer to documentation for bowtie options:
-
-
--bowtie_build_cmd - command to run bowtie build, must be available in PATH or resolve to bowtie binary (default: "bowtie2-build")
-
--bowtie_build_args.[argument name (with - or -- if required)]=[argument value] - arguments to be passed to bowtie build command. defaults:
-
-f=[path to fasta file]
-
refer to documentation for bowtie index build options:
-
-
--index_file - path to save index files to (default: temp file)
-
--fastq_file_list - csv file of fastq files to run (optional, files will be filtered to those matching files included in this list, prefixes are permissable)
-
--remove_index - remove index files after running (default: true)
-
--output_format - format to save count and summary files, json or csv (default: csv)
-
--concurrency - number of alignments / counts to run at once (default: 1)
-
--help|h|usage - print usage
-
--verbose - print child process output (default: false)
==================
count bowtie reads across multiple samples and aggregate into a csv file
================== required arguments:
- --reads_list - csv file including a list of reads to aggregate (format: single column, titled "read")
- --count_files_list - csv file including list of samples to include in aggregated counts file (format: column titled "sample" [required] with name of sample [lane is optional, multiple lanes of the same sample will be consolidated in output file], column titled "name" [optional] with a user-specified name for the sample)
- --count_files_path - path to directory containing .csv or .json files with bowtie read counts (generated by bowtie-counter script)
- --output_path - path to save csv file of aggregated counts to, (format: column titled "read" includes each read, columns for each sample [labelled with user-supplied name - if provided - and file name of sample], containing number of times each read appeared in the sample)
================== optional arguments:
- --concurrency - number of count files to tabulate at once (default: number of CPU cores on current system)
- --help|h|usage - print usage
==================
These scripts are currently designed to run with Node.js, which is ideal for high-thoroughput and parallelizable disk / process IO required for bowtie alignments. The scripts make minimal use of temporary files, and pipe process output directly, speeding up the counting / aggregation process. They have been tested on Linux, but should work find on Windows and Mac sytems. Additionally, they have been tested running on up to 40 parallel CPUs on Amazon EC2 (m4.10xlarge) instances.