lisavader/plasmidec

An ensemble of plasmid classification tools

plasmidec's Issues

Problem with Argument parsing

There is some problem with the way the arguments are handled.
If I don't give -g before -t the $threads will be null and $SCRIPT_DIR will not be parsed into the subscripts

while getopts :i:c:o:fgtvh flag; do

Changed to

while getopts ":i:c:o:fgt:vh:" flag; do

Now working fine but still needs your review.

centrifuge build version

Hi, Julian and I spent the better part of today trying to figure out why his plascope did something different compared to mine. The outcome: I had an older build version of centrifuge.

Proposed solution: hardcode the build version of centrifuge in the command installing plascope/centrifuge in the plasmidEC_plascope environment, or (in the case this is handled by plascope) update it after installing plascope.

Code that solved my issue:
conda install centrifuge=1.0.3=py36pl5.22.0_3 -c bioconda

Problem with plasmid_contigs.fasta

The contigs are not complete

I added seqtk to get the fasta and I wanted the file name in the contigs. So add these things in the write_plasmid_contigs.sh

for contig in $plasmid_contigs; do
        echo $contig >> $out_dir/contig.txt
        #grep -A 1 $contig $input >> $out_dir/plasmid_contigs.fasta
done

TAG="$(basename $input .fasta)"
seqtk subseq $input $out_dir/contig.txt >> $out_dir/"${TAG}"_plasmids.fasta
sed -i "s/^>/>${TAG}_/" $out_dir/"${TAG}"_plasmids.fasta

Because the -A 1 in grep does grep only subsequent line after the header.

-A, --after-context=NUM   print NUM lines of trailing context

The contigs from spades are never in a single line.

Plascope - Centrifuge error

Centrifuge does not find the correct perl installation in the HPC. Therefore Plascope does not produce plasmid predictions. See error below:
Cwd.c: loadable library and perl binaries are mismatched (got handshake key 0xde00080, needed 0xdb00080)

Temporary fix:
Modify the PERL5LIB path in .bashrc file. Add following line:
export PERL5LIB=/hpc/dla_mm/jpaganini/data/miniconda3/envs/plascope_ec_lv/lib/perl5/5.22.0/
This directs centrifuge to the correct perl libraries in the plascope_ec_ev (the one created for running the plascope within the EC)
(Each user will have to modify their own path to the miniconda or conda installation)

Missing R packages

Hi! I'm running plasmidEC to work with gplas, and it seems that two R packages are not (correctly) installed in the plasmidEC_R conda environment:

• plyr
• dplyr

Installing them fixes the issue, but it would be nice if plasmidEC did this while making the environment

Update:
It might be due to an earlier error with R where some C libraries couldn't be found. This was fixed by removing the conda env and recreating the env. Might be a problem on our HPC.

Wish list

Wish list

• getting the gplas-formatted output should be optional, with a flag.
• multi-threading the individual tools, if possible
• input as file instead of directory
• fasta file with all the plasmid sequences
• provide tools selection as default. Optional flag to choose different tools.
• update readme + add help flag
• add version