nextstrain

Overview

The automated_nextstrain.sh script automates the daily execution of TGen's Nextstrain analysis pipeline. This script is configured to run as a cron job, processing GISAID data, sampling genomes, and building a Nextstrain tree for Arizona-specific sequences.

SLURM Directives

The script begins with SLURM job specifications:

#SBATCH --cpus-per-task=14
#SBATCH --mem=140G
#SBATCH --partition=data-mover

• CPUs: Allocates 14 CPUs for the task.
• Memory: Allocates 140 GB of RAM.
• Partition: Runs on the data-mover partition.

---

Environment Setup

source /home/bvan-tassel/miniforge3/etc/profile.d/conda.sh
conda activate R

The script activates the R Conda environment, used later for genome sampling and other R-based tasks.

---

Directory Preparation

rm -r running_NS
cp -R ${builder} running_NS
cp -R ${config_alt}/${style}/config.json running_NS/config/config.json

• Removes any existing running_NS directory.
• Copies the nextstrain_weekly builder (containing the Snakemake workflow) into the running_NS directory.
• Updates the Nextstrain configuration file with Arizona-specific settings.

---

Genome Sampling

Rscript sample_gisaid.R ${style}

This command uses an R script to sample genomes from the GISAID dataset based on a likelihood model, ensuring newer genomes are more likely to be included in the analysis.

---

Color Assignment

conda run -n nextstrain python auto_color_assign.py

Assigns colors to new lineages for tree visualization in Nextstrain. The color information is added to the configuration folder.

---

Tree Rooting

rooter=$(cat root.name)
sed -i -e "s/VttLJ2gz1X4B/${rooter:-oldest}/" running_NS/Snakefile

• Retrieves the root name from the root.name file.
• Updates the Snakemake file to use the correct root for the phylogenetic tree.

---

Linking Data Files

ln -s $( readlink -e subset_sequences.fasta ) running_NS/data/sequences.fasta
ln -s $( readlink -e subset_metadata.tsv ) running_NS/data/metadata.tsv

Creates symbolic links for the sampled sequences and metadata files, making them accessible to the Snakemake workflow.

---

Running the Nextstrain Workflow

conda activate nextstrain
export AUGUR_RECURSION_LIMIT=10000
cd running_NS
snakemake --cores 14

• Activates the nextstrain environment.
• Sets AUGUR_RECURSION_LIMIT to handle large datasets.
• Executes the Snakemake workflow using 14 cores.

---

GitHub Update

python3 github_upload.py

Uploads the results to a GitHub repository. This includes metadata and color files used in the Nextstrain visualization.

---

Cron Job Integration

The script is scheduled to run daily using a cron job. Example cron entry:

40 1 * * * source /home/bvan-tassel/miniforge3/etc/profile.d/conda.sh && conda activate nextstrain; cd /tnorth_labs/COVIDseq/nextstrain/ && sh automated_nextstrain.sh | mail -s "Nextstrain update" [email protected]

This schedules the script to run at 1:40 AM every day.

---

Script Location

Ensure the script is stored in a directory accessible to the SLURM environment and cron job.

---

Dependencies

1. Conda Environments:
   • R
   • nextstrain
2. External Tools:
   • Snakemake
   • AUGUR

Output

The workflow generates:

• A phylogenetic tree in Nextstrain format.
• Updated metadata and lineage colors.
• Results pushed to a designated GitHub repository.

https://nextstrain.org/community/tgennorth/arizona-covid-19/AZ?d=tree&onlyPanels&p=full&sidebar=closed

Notes

• Ensure SLURM, Conda, and Python configurations are properly set up on your system.
• Verify the paths to GISAID data files (seq and meta) and configuration files.