This repository contains a (snakemake) workflow to build Metagraph indexes and bash scripts to run queries against the indexes that have been used for the experiments in this paper https://doi.org/10.1101/2023.07.21.550101.
snakemake
v7.30 (recommended installation viaconda
);metagraph
, commit 5c2a12b.
snakemake --cores all --configfile <config.yml> -- <rule>
Rules to produce various annotations:
row_diff
row_diff_flat
row_diff_sparse
row_diff_brwt
relax_brwt_arity
For example:
snakemake --cores 48 --configfile se-150k.yml -- relax_brwt_arity
Note: building any of the variants produces the "plain" variant that is not row-diff compressed.
Use provided python script to sum build times for metagraph variants:
python gather_build_times.py <path_to_log_dir> <mode>
Logs are written to <output-dir>/<exp-name>/logs
.
- outputs are written to
<output-dir>/<exp-name>
withoutput-dir
defaulting tooutput/<exp-name>
if not specified. - file containing list of filepaths is specified with
input-file-lists
- path to
metagraph-bin
metagraph binary must be supplied. - Number of cores used set to
threads
where possible.
See the file example-config.yml
for an example.
See the examples provided in the scripts high_hit_queries.sh
and low_hit_queries.sh
.