egaffo / circompara2

Improved bioinformatic pipeline to identify and quantify circRNA expression from RNA-seq data by combining multiple circRNA detection methods
Other
7 stars 0 forks source link

Update mappers #18

Open jaanckae opened 1 year ago

jaanckae commented 1 year ago

Some of the software dependencies in the Docker container are several versions behind a recent version. Would it be possible to update the Dockerfile to work with these updated versions?

At least for some versions of the mappers this would be convenient as not to have to rerun the indexing script.

Software Website used_version latest_release
Scons http://www.scons.org 3.1.2 4.5.2
Trimmomatic http://www.usadellab.org/cms/?page=trimmomatic 0.39 0.39
FASTQC http://www.bioinformatics.babraham.ac.uk/projects/fastqc/ 0.11.9 0.12.1
HISAT2 http://ccb.jhu.edu/software/hisat2/index.shtml 2.1.0 2.2.1
STAR http://github.com/alexdobin/STAR 2.6.1e 2.7.11
BWA http://bio-bwa.sourceforge.net/ 0.7.15-r1140 0.7.17
Bowtie2 http://bowtie-bio.sourceforge.net/bowtie2/index.shtml 2.4.1 2.5.1
Bowtie http://bowtie-bio.sourceforge.net/index.shtml 1.1.2 1.3.1
TopHat http://ccb.jhu.edu/software/tophat/index.shtml 2.1.0 2.1.1
Segemehl http://www.bioinf.uni-leipzig.de/Software/segemehl/ 0.3.4 0.3.4
CIRI http://ciri.sourceforge.io/ 2.0.6 2.1.1
CIRCexplorer2 http://github.com/YangLab/CIRCexplorer2 2.3.8 2.3.8
find_circ http://github.com/marvin-jens/find_circ 1.2 1.2
BEDtools http://bedtools.readthedocs.io 2.29.2 2.31.0
Samtools http://www.htslib.org/ 1.10 1.18

Thanks in advance Jasper

egaffo commented 1 year ago

Some packages must stay at the version currently implemented in CirComPara2 because of incompatibility issues. F.i. I achieved TopHat2 and Bowtie to work together only with exactly those two versions! In some other tool newer versions, I found bugs that are still unresolved (see f.i. HISAT2 #255). Further, newer versions did not provide improvements for circRNA detection or computation speed (BEDtools, SAMtools, BWA, FASTQC), and CIRI v2.1.1 refers to the CIRI_full algorithm, which actually uses CIRI v2.0.6. Then, I see no reason to spend time updating these packages. Perhaps STAR and Bowtie2 newer versions could give some improvements (as I understand from the release statements), but STAR changed the Chimeric.junction file format that would break the STAR-based detection tools, i.e. circRNA_finder, CIRCexplorer2, and DCC v0.4.8 (DCC v0.5 was indeed released to fix this issue). Unfortunately, you'll have to use a STAR index compatible with v2.6.1e . Making all these tools work together was worse than playing Tetris level 29.