K2Rmini Experiments

Experiments on K2Rmini and other sequence filtering tools.

Requirements

Make sure to install the latest Rust version with rustup.

You'll also need a Python setup with pandas, matplotlib and seaborn to make the plots, and different development librairies depending on your OS.

Debian-based distributions

sudo apt install -y libsqlite3-dev libzstd-dev libhyperscan5 libhyperscan-dev pipx python3-pandas python3-matplotlib python3-seaborn

Fedora-based distributions

sudo dnf install -y libsq3-devel libzstd-devel hyperscan-devel pipx python3-pandas python3-matplotlib python3-seaborn

macOS

brew install gnu-time zstd vectorscan pipx

Installation

You can install all the tools using

bash install.sh

This will store all the binaries in the bin folder.

Running experiments

You can run experiments on your desired set of reads (in FASTQ format). For instance, you can use these PacBio HiFi reads from HG002.

You can adjust the number of repetitions (-R), timeout for each command (-T), number of threads used by each tool (j), k-mer size (-k) and the directories used to store logs (-l) and patterns (-p). A detailed description of each flag is available using -h.

python3 experiments.py -r <READS> -R 5 -T 300 -j 8

This script will run multiple tools and store the results of each run in the log folder.

Plotting the results

Once some results are available in the log folder, you can generate plots comparing the runtime and memory usage of the different tools.

You can change the output format (-f), tools shown (-s) or excluded (-x) from the plots, and the directories used for the logs (-l) and plots (-p).

Three types of plots are available:

• time/memory vs number of patterns --versus num_patterns (default)
• time/memory vs number of threads using --versus threads
• time/memory vs k-mer size using --versus k

python3 plot.py -f pdf png

Experiments from the paper

# varying number of patterns
python3 experiments.py -r <READS> -R 5 -T 300 -k 31 -j 1 8 -l log
python3 plot.py -f pdf png --versus num_patterns -l log

# varying number of threads
python3 experiments.py -r <READS> -R 5 -T 300 -k 31 -n 1048576 -j 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 -l logt -s k2rmini deacon sbwt bts cleanifier
python3 plot.py -f pdf png --versus threads -l logt

# varying k-mer size
python3 experiments.py -r <READS> -R 5 -T 300 -j 1 -n 1048576 -k 31 39 47 55 63 71 79 87 -l logk -s k2rmini deacon sbwt bts
python3 plot.py -f pdf png --versus k -l logk