Scanpy

Introduction

Scanpy is scalable toolkit for analyzing single-cell gene expression data. It includes preprocessing, visualization, clustering, pseudotime and trajectory inference and differential expression testing. The Python-based implementation efficiently deals with datasets of more than one million cells. Details about its usage can be found here (https://scanpy.readthedocs.io/en/stable/)

Versions

1.8.2
1.9.1

Commands

python
python3

Module

You can load the modules by:

module load biocontainers
module load scanpy/1.8.2

Interactive job

To run scanpy interactively on our clusters:

(base) UserID@bell-fe00:~ $ sinteractive -N1 -n12 -t4:00:00 -A myallocation
salloc: Granted job allocation 12345869
salloc: Waiting for resource configuration
salloc: Nodes bell-a008 are ready for job
(base) UserID@bell-a008:~ $ module load biocontainers scanpy/1.8.2
(base) UserID@bell-a008:~ $ python
Python 3.9.5 (default, Jun  4 2021, 12:28:51)
[GCC 7.5.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import scanpy as sc
>>> sc.tl.umap(adata, **tool_params)

Batch job

Warning

Using #!/bin/sh -l as shebang in the slurm job script will cause the failure of some biocontainer modules. Please use #!/bin/bash instead.

To submit a sbatch job on our clusters:

#!/bin/bash
#SBATCH -A myallocation     # Allocation name
#SBATCH -t 10:00:00
#SBATCH -N 1
#SBATCH -n 24
#SBATCH --job-name=scanpy
#SBATCH --mail-type=FAIL,BEGIN,END
#SBATCH --error=%x-%J-%u.err
#SBATCH --output=%x-%J-%u.out

module --force purge
ml biocontainers scanpy/1.8.2

python script.py