Download - Oshlack/JAFFA GitHub Wiki
Latest JAFFA release
Download the latest (or older) releases from the github release page:
Reference files
You will also need one of these supporting reference files:
- JAFFA_REFERENCE_FILES_HG38_GENCODE49.big.tar.gz - Files for hg38 (JAFFA 2.5+ assumes this reference by default)
- JAFFA_REFERENCE_FILES_HG38_GENCODE22.V2.tar.gz - Files for hg38 (JAFFA 1.07-2.4 assumes this reference by default)
T2T, hg19, mm39 and mm10 are also supported, but require you to follow the instructions here in addition to downloading the reference below
- JAFFA_REFERENCE_FILES_HS1_CATLIFTOFFGENESV1.big.tar.gz - Files for human T2T (JAFFA 2.5+ only)
- JAFFA_REFERENCE_FILES_HG19_GENCODE19.tar.gz - Files for human hg19 (JAFFA v2.4 and below)
- JAFFA_REFERENCE_FILES_MM39_GENCODEVM38.big - Files for mouse mm39 (All JAFFA versions)
- JAFFA_REFERENCE_FILES_MM10_GENCODEVM4.tar.gz - Files for mouse mm10 (JAFFA 2.4 and below only)
Please see FAQandTroubleshooting if you would like to generate your own reference files for JAFFA (e.g. for an unsupported genome or annotations).
If you are finding the downloads above too big, from JAFFA v2.5 onwards we provide a script within JAFFA, prepare_reference.sh, which will automatically fetches a smaller minimum set of files and generates all other required reference files for hg39, T2T or mm39. The script should be run after running install_linux64.sh, or equivalently setting up all the required software and specifying their path in tools.groovy. The script will take a few hours to be run, but only needs to be performed once.
Datasets
Demo data - See Example for running the demo
Simulations from our papers:
- BEERS (no fusions) - BEERS_1.fastq.gz BEERS_2.fastq.gz
- MiSeq (120 fusions) - MiSeqSim_1.fastq.gz MiSeqSim_2.fastq.gz MiSeqSim_truth.tar.gz
- ONT and PacBio (2500 fusions) - LongReadFusionSimulation.tar.gz
Versions prior to 1.06 that are not hosted on github
JAFFA_1.05.tar.gz - Fusion gene names are now ordered by fusion gene orientation (previously gene names were listed alphabetically)
JAFFA_1.04.tar.gz - This version fixes an issue that caused the direct mode to be slow on some large dataset.
JAFFA_1.03.tar.gz - Major improvements over 1.0 such as:
- Reads that map to intronic, intergenic, and chrM regions are now first filtered out before running the pipeline. This improves the RAM and time requirements for the assembly and hybrid modes.
- For the direct mode we now calculate spanning pairs
- fusions classed as PotentialRegularTranscripts are no longer reported
- we now threshold on 2 spanning reads for LowConfidence fusions (previously this was 1), to help remove library prep false positives.
JAFFA_1.0.tar.gz -This version is identical to 0.91. Improvements of 0.9 are that it works out the read length itself and includes a few bug fixes and checks. Thanks to Aliaksei Holik for correcting the bug in the spanning pairs calculation.