So, when asking my 'peeps' and looking at the literature on this process, it appears that dDocent is the most favored approach for assembling ddRADSeq data. As it turns out, this is mostly just a wrapper around a wide variety of additional (existing) tools and all this one does is coordiante them individually..
Here is a list of all the stuff that I needed to install on my cluster to get it going.
To Install The following items have yet to be installed. I don't know if I'll actually need this.
- java
Installed I've installed the following packages (in this order):
- STACKS
- FreeBayes
- Trimmomatic
- BWA: See the manual pages for the Burrows-Wheeler Alignment Tool.
- VCFtools
- rainbow
- seqtk
- CD-HIT
- BEDOPS
- bedtools
- bamtools
- Mawk: A AWK interpreter.
- PEAR read merger: I actually installed this and it installed
pearbut he dDocent script was looking forpearRM. I set up a softlink for it.
Modules The following were modules module load xxx on the server. For some reason we have some kinds of things set up as modeules that need to be loaded each time. These are the ones that are already on the cluster.
$ module load R/4.1.2
$ module load samtools/1.12
$ module load bcftools/1.12
$ module load htslib/1.12
Conda Modules The following are anaconda modules (and I'm waiting to see if I need to get my own python installed because the one on huff is out-of-date). To get this to work, I had to install pyenv because the system admins had version 2.9 (latest is 3.11 I think) and none of the stuff would compile on it as it stands.
- pyenv Gotta install this so I can install a version of python that is up-to-date. Intalled this and have anaconda3 set as the default local version. I put the init scripts to this in my .bash_provile so that it loads with each log in and I don't have to futz around with geting it up and going each time something is needed to run in python. Apparently a lot of conda install ... crap is needed.
May Not Need The following stuff may not be needed
- gnuplot: Already installed.
- gnu-parallel.