Performs merging and scoring from multiple mapped transposon experiments.
Requires Java and commandline access to run. Download the latest release jar: hbc.transposon-0.0.4-standalone.jar (25.3Mb).
Runs in a two step process. The first takes a YAML config or Excel sample file and merges into a single output CSV file:
$ java -jar hbc.transposon-0.0.4-standalone.jar merge <work_directory> -c <YAML config file> -x <Excel sample file>
The second scores and filters the merged samples:
$ java -jar hbc.transposon-0.0.4-standalone.jar score <merged CSV file> -c <YAML config file> -x <Excel sample file>
To control filtering of contamination two options are available:
-f 50Specify a hard cutoff of read counts to filter below.
-p 0.95Specify a percentile to filter below. The algorithm calculates read count from the distribution of potential contamination.
Written in Clojure and the requires the leiningen build tool to run directly from source:
$ lein merge <work_directory> -c <YAML config file> -x <Excel sample file> $ lein score <merged CSV file> -c <YAML config file> -x <Excel sample file>
We're revisiting this project using UMI tagged transposons sequences to identify duplicates in the analysis. The workflow for starting from raw fastq files.
Overlap paired end reads and extract genomic sequence for the insertion site and UMI barcodes, creating a fastq file for each:
python scripts/prepare_umis.py <fastq read 1> <fastq read 2>
Using genomic fastq as single end data and UMIs as barcodes (
umi_type), run an alignment with bcbio (https://github.com/chapmanb/bcbio-nextgen)
Convert UMI tagged BAMs into unique genome positions:
python scripts/bams_to_insertionsites.py <BAM file 1> <BAM file 2>