e-coli

E.coli experiment with natural reads and known variants

This directory contains the scripts needed to run the E.coli experiments with natural reads.

Running the experiment

1. Download the indexing input with e.g. wget https://cs.helsinki.fi/group/gsa/panvc-founders/natural-e-coli-experiment/e-coli-index-input.tar.bz2. Extract the contents to the e-coli directory with e.g. pbzip2 -dc e-coli-index-input.tar.bz2 | tar x.
2. Download the reads used in the experiment with e.g. wget https://cs.helsinki.fi/group/gsa/panvc-founders/natural-e-coli-experiment/e-coli-reads.tar. Extract the contents with tar x e-coli-reads.tar.
3. Generate an index using the founder sequences with snakemake --configfile config-common-index.yaml config-index/e-coli-msd5.yaml --snakefile ../panvc-sample-workflow/Snakefile.index --cores 16 --printshellcmds --use-conda --conda-prefix ../conda-env --resources mem_mb=16384.
4. Generate an index using the random selection of reference sequences with snakemake --configfile config-common-index.yaml config-index/e-coli-subset20.yaml --snakefile ../panvc-sample-workflow/Snakefile.index --cores 16 --printshellcmds --use-conda --conda-prefix ../conda-env --resources mem_mb=16384.
5. The experiments may be run with e.g. ./founder-experiment-cmds.sh | parallel -jjobs where jobs is the number of Snakemake instances. The script allocates 16 GB of memory and 5 CPU cores to each instance of Snakemake. Alternatively the commands may be piped to e.g. bash: ./founder-experiment-cmds.sh | bash -x -e.
6. The experiments with the random selection of reference sequences may be run similarly with either ./subset20-experiment-cmds.sh | parallel -jjobs or ./subset20-experiment-cmds.sh | bash -x -e.
7. To gather statistics with Samtools, run first mkdir samtools-stats and then ./run-samtools-stats.sh. The results will be placed in the aforementioned directory.

Running the experiment with the comparison to de novo sequenced contigs

1. Download the reads used in the experiment with e.g. wget https://cs.helsinki.fi/group/gsa/panvc-founders/natural-e-coli-experiment/e-coli-reads-de-novo-experiment.tar. Extract the contents to the e-coli directory with tar x e-coli-reads-de-novo-experiment.tar.
2. Download the contigs used in the experiment with e.g. wget https://cs.helsinki.fi/group/gsa/panvc-founders/natural-e-coli-experiment/e-coli-de-novo-contigs.tar. Extract the contents with tar x e-coli-de-novo-contigs.tar.
3. Download the E.coli K-12 reference without line breaks with e.g. wget https://cs.helsinki.fi/group/gsa/panvc-founders/natural-e-coli-experiment/e.coli-k12-mg1655.no-linebreaks.fa.gz. Extract with gunzip e.coli-k12-mg1655.no-linebreaks.fa.gz.
4. The indices generated in the previous experiment may be used.
5. Run the following commands where jobs is the number of Snakemake instances.
   • ./founder-de-novo-experiment-cmds.sh | parallel -jjobs
   • ./subset20-de-novo-experiment-cmds.sh | parallel -jjobs
6. Generate the predicted sequences with snakemake --printshellcmds --use-conda --conda-prefix ../conda-env. Here Snakemake is only used to activate the correct Conda environment; the snakefile essentially runs predicted-sequences.sh.
7. Run QUAST with mkdir -p quast && ./quast.sh. The results will be placed in quast.