Saturn is a deep learning approach that couples gene expression with protein representations learnt using large protein language models for cross-species integration. The key idea in SATURN is to map cells from all datasets to a shared space of functionally related genes that we name macrogenes. Using macrogenes, SATURN is uniquely able to detect functionally related genes co-expressed across species.

References:

• Rosen Y, Brbić M, Roohani Y, Swanson K, Li Z, Leskovec J. Toward universal cell embeddings: integrating single-cell RNA-seq datasets across species with SATURN. Nat Methods. 2024 Aug;21(8):1492-1500. doi: 10.1038/s41592-024-02191-z. Epub 2024 Feb 16. PMID: 38366243; PMCID: PMC11310084. PubMed |  Journal

• saturn Main Site

• Usage: python-saturn

[user@helix]$ module load saturn
[user@helix]$ python-saturn

Allocate an interactive session and run the program.
Sample session (user input in bold):

[user@biowulf]$ sinteractive --gres=lscratch:30,gpu:a100:1 -c 8 --mem=60g  
salloc.exe: Pending job allocation 46116226
salloc.exe: job 46116226 queued and waiting for resources
salloc.exe: job 46116226 has been allocated resources
salloc.exe: Granted job allocation 46116226
salloc.exe: Waiting for resource configuration
salloc.exe: Nodes cn3144 are ready for job

[user@cn3144 ~]$ module load saturn

[user@cn3144 ~]$ cd /data/$USER/

[user@cn3144 ~]$ python-saturn

Create a batch input file (e.g. saturn.sh). For example:

#!/bin/bash
set -e
module load saturn
python-saturn ../SATURN/train-saturn.py xxxx

Submit this job using the Slurm sbatch command.

sbatch [--cpus-per-task=#] [--mem=#] saturn.sh