RNApeg is an RNA junction calling, correction, and quality-control package. Its use is discussed in this paper:
https://rnajournal.cshlp.org/content/24/8/1056.short
Invoke the RNApeg wrapper as
RNApeg.sh [-h] -b bamfile -r refflat -f fasta
Where
- -b is a BAM file of transcriptome reads mapped to whole-genome coordinates
- -r is an uncompressed refFlat.txt file from the UCSC genome annotation database
- -f is a FASTA file with a .fai-format index (generated by "samtools faidx genome.fa")
See 'Downloading reference files' to learn how you can obtain the most up to date reference files.
Output is written in the Docker container to the directory /results. This path must be mounted at runtime, i.e. either to a volume or bind-mounted to a directory on your local computer.
The following example:
- mounts the volume "myvolume" to "/results", where output file will be written
- bind-mounts the local PC directory /c/Users/myuid/docker_rnapeg to /mny/mydata in the counter
- specifies refFlat and genome files stored on the PC
docker run --mount source=myvolume,target=/results -v /c/Users/myuid/docker_rnapeg:/mnt/mydata rnapeg RNApeg.sh -g GRCh37-lite -b /mnt/mydata/demo/rna_demo_TP53.bam -r /mnt/mydata/refFlat.txt -f /mnt/mydata/reference/Homo_sapiens/GRCh37-lite/FASTA/GRCh37-lite.fa
- This software was originally written to run in a cluster computing environment, and has substantial RAM requirements (~8+ gb, still TBD). On a PC, the Docker virtual machine memory allocation almost certainly needs to be increased from the default values. Use of a computer with 16 gb of RAM is recommended, an 8 gb system may not be sufficient.
- GNU parallel
- Picard 2.6.0
- Java 1.8.0
- Perl 5.10.1 with libraries:
The UCSC genome annotation database provides genomic mappings of NCBI refSeq transcripts for various genomes. The refFlat file for hg19 is:
http://hgdownload.cse.ucsc.edu/goldenPath/hg19/database/refFlat.txt.gz
This file must be decompressed before use.
- provide copy of RNApeg wiki docs?