####log into your account on SuperMikeII ####go to your work directory on the cluster cd /work/albuseb #or whatever it is
####make a directory to store the FASTA files for DeconSeq mkdir deconseq_ref #or whatever you want to call the reference directory
####make a directory to store the output files for DeconSeq mkdir deconseq_out
####go to the deconseq_ref directory that you just made cd deconseq_ref
####get all virus FASTA files (here fna is for Fasta Nucleic Acid) wget ftp://ftp.ncbi.nih.gov/genomes/Viruses/all.fna.tar.gz
####rename the file to all.vir.fna.tar.gz mv all.fna.tar.gz allvir.fna.tar.gz
####extracts only the fna files and not the folders (the strip-components part) from all.fna.tar.gz tar -xzf allvir.fna.tar.gz --strip-components 2
####creates a FASTA file called virus.fa from all the .fna files (1 after the other, after the other) cat *.fna > virus.fa
####gets rid of all of the .fna files (you don't need them any more rm *.fna
####repeat above steps for bacteria ####get all bacteria FASTA files, extract only .fna files (might take 3-5 minutes), combine into one fasta file (take 1-2 minutes), get rid of .fna files wget ftp://ftp.ncbi.nih.gov/genomes/Bacteria/all.fna.tar.gz mv all.fna.tar.gz allbac.fna.tar.gz tar -xzf allbac.fna.tar.gz --strip-components 2 cat *.fna > bacteria.fa rm *.fna
####combine virus.fa and bacteria.fa cat bacteria.fa virus.fa > deconseq_ref.fa
####Now you need to install DeconSeq ####go back to your home folder cd /home/albuseb
####make a new directory called bin to store your program in mkdir bin
####make a new directory called scripts to store your scritps in mkdir scripts
####go to this directory cd scripts
####make new directory for deconseq scripts mkdir deconseq
####go to this directory cd deconseq
####you need to put the following scripts here (not sure how you are doing file transfers) #####make_BWA_index.sh #####splitFasta.sh wget https://raw.githubusercontent.com/jelber2/deconseq/master/make_BWA_index.sh wget https://raw.githubusercontent.com/jelber2/deconseq/master/splitFasta.sh
#####you need to change the allocation for each file using the text editor nano nano make_BWA_index.sh #####find #PBS -A hpc_startup_albuseb #####change hpc_startup_albuseb to your allocation name #####once done making changes, press 'ctrlX' (ctrl plus X) ######you will see Save modified buffer (ANSWERING "No" WILL DESTROY CHANGES) ? ######then press 'Y' to save changes ######you will then see File Name to Write: make_BWA_index.sh ######press 'Enter' to save file as itself #####do the same thing for splitFasta.sh nano splitFasta.sh
####go back to bin folder cd .. cd .. cd bin
####download DeconSeq (deconseq-standalone-0.4.3.tar.gz) to bin folder wget http://downloads.sourceforge.net/project/deconseq/standalone/deconseq-standalone-0.4.3.tar.gz?r=http%3A%2F%2Fsourceforge.net%2Fprojects%2Fdeconseq%2Ffiles%2Fstandalone%2F&ts=1417735856&use_mirror=hivelocity
####unzip the compressed tar file tar -xzf deconseq-standalone-0.4.3.tar.gz
####go to the new directory cd deconseq-standalone-0.4.3
####use splitFasta.pl to split the 10GB file into 10 files ####note the command below must be implemented with splitFasta.sh (because LSU HPC does not like processes running on the head node) ####perl splitFasta.pl -verbose -i /work/albuseb/deconseq_ref/deconseq_ref.fa -n 10
####to use splitFasta.sh, enter the following in the terminal qsub /home/albuseb/scripts/deconseq/splitFasta.sh
####check the status of the job using qstat -u albuseb
####combine the small files into a larger one cd /work/albuseb/deconseq_ref cat deconseq_ref.fa_c4.fasta deconseq_ref.fa_c5.fasta deconseq_ref.fa_c6.fasta deconseq_ref.fa_c7.fasta deconseq_ref.fa_c8.fasta deconseq_ref.fa_c9.fasta deconseq_ref.fa_c10.fasta > deconseq_ref.fa_c4a.fasta
========
####Time to make BWA indexes ####assumes you are in /home/albuseb/bin/deconseq-standalone-0.4.3/ ####make bwa executable chmod u+x bwa64
####test bwa64 ./bwa64
####should see: Program: bwa (alignment via Burrows-Wheeler transformation) Version: 0.5.9-r16 Contact: Heng Li [email protected] etc...
####use make_BWA_index.sh to implement the following commands: /home/albuseb/bin/deconseq-standalone-0.4.3/bwa64 index -a bwtsw -p deconseq_ref.fa_c1 deconseq_ref.fa_c1.fasta /home/albuseb/bin/deconseq-standalone-0.4.3/bwa64 index -a bwtsw -p deconseq_ref.fa_c2 deconseq_ref.fa_c2.fasta etc.
####to use make_BWA_index.sh qsub /home/albuseb/scripts/deconseq/make_BWA_index.sh
####change the location of the database(s) in the DeconSeqConfig.pm file nano DeconSeqConfig.pm
####scroll to use constant DB_DIR => 'db/'; #####change to '/work/albuseb/deconseq_ref/
####scroll to use constant DBS => #####change hsref to deconseq_ref #####change 'Human Reference GRCh37' to 'NCBI_ALL_BAC_&_Vir' #####change 'hs_ref_GRCh37' to 'deconseq_ref.fa_c1,deconseq_ref.fa_c2,deconseq_ref.fa_c3,deconseq_ref.fa_c4a' ####scroll to use constant DB_DEFAULT => #####change to hsref to deconseq_ref
========
###you are now ready to run DeconSeq ####first you need 'edit' deconseq.pl so it will run faster using 16 rather than 1 core #####all I had to do was add '-t 16' to two lines in the perl script to make #####deconseq_16cores.pl #####use wget to download the 16 core version from github #####assumes you are in /home/albuseb/bin/deconseq-standalone-0.4.3/ directory wget https://raw.githubusercontent.com/jelber2/deconseq/master/deconseq_16cores.pl
####now get the run_deconseq.sh script so you can run the program on a compute node cd /home/albuseb/scripts/deconseq wget https://raw.githubusercontent.com/jelber2/deconseq/master/run_deconseq.sh
####this is the code the script implements
cd /home/albuseb/bin/deconseq-standalone-0.4.3/
perl /home/albuseb/bin/deconseq-standalone-0.4.3/deconseq_16cores.pl
-f /work/albuseb/folder_containing_fastq_or_fasta_to_filter/file_to_filter.fastq
-dbs deconseq_ref
-out_dir /work/albuseb/deconseq_out
####now you need to edit the script to change it to your allocation and also path to files to filter using nano #####you may also need to change the options (right now these are default settings for DeconSeq) nano run_deconseq.sh
####you can submit the job using qsub /home/albuseb/scripts/deconseq/run_deconseq.sh
####check it's status with qstat qstat -u albuseb