Guppy
Description
Guppy is a data processing toolkit that contains the Oxford Nanopore Technologies' basecalling algorithms, and several bioinformatic post-processing features.
License
General Nanopore Product Terms and Conditions
To be able to use the application, you need first to accept its licence
WARNING: Please be aware that guppy-*-gpu versions require nodes with GPUs.
GPU versions utilise the Cuda toolkit and you have to accept a license for cuDNN library.
Usage
Upcoming modulesystem change alert!
Due to large number of applications and their versions it is not practical to keep them explicitly listed at our wiki pages. Therefore an upgrade of modulefiles is underway. A feature of this upgrade will be the existence of default module for every application. This default choice does not need version number and it will load some (usually latest) version.
You can test the new version now by adding a line
source /cvmfs/software.metacentrum.cz/modulefiles/5.1.0/loadmodules
to your script before loading a module. Then, you can list all versions of guppy and load default version of guppy as
module avail guppy/ # list available modules module load guppy # load (default) module
If you wish to keep up to the current system, it is still possible. Simply list all modules by
module avail guppy
and choose explicit version you want to use.
guppy_basecaller --help
Guppy basecalling can be significantly accelerated on GPU clusters. For the utilization of GPU mode, you have to use the appropriate Guppy module with the -gpu extension and agree with Guppy and cuDNN licenses (see above).
The example below will use one GPU card and consume ~38 GB of GPU physical memory.
module add guppy-6.0.6-gpu guppy_basecaller -i input_fast5 -r -s out_fastq_reads --flowcell FLO-MIN106 --kit SQK-LSK109 -x auto --gpu_runners_per_device 16 --num_callers 16 --chunks_per_runner 2000 --trim_strategy none --disable_qscore_filtering
Physical GPU memory does not work as a PBS parameter :mem= and cannot be controlled. Reducing their values will reduce the consumed GPU memory (ie. the calculation can be run on GPUs with smaller memories), but the calculation will be slower. Better performance can be achieved by using scratch on fast SSD discs :scratch_ssd=.