Perform exhaustive short read mapping against a reference database.
Specifically, we generate all successive k length substrings, jumping j
characters, across the set of input sequences. These substrings are then
mapped against a database to determine regions of high identity. Mapped
coordinates are transformed into contiguous regions, and a fasta file
containing the observed sequences is produced. Finally, the database is masked
such that regions of similarity are converted to N characters, and a new
bowtie2 index is then constructed.
WYSIWYG for the time being. The scripts are designed around a SLURM submission system. The included Python scripts should generalize across Python versions with the only dependencies being Click and NumPy.
submit-exhaustive.sh requires a label, the database as fasta or xz compressed
fasta, a bowtie2 index to map against, and a file containing the paths to the
sequence files to generate short reads from. The submission will issue calls
to sbatch and setup dependencies for subsequent work.