First Analysis¶

This guide walks through analyzing your own aa-tRNA-seq data.

Prerequisites¶

• Completed Installation
• POD5 files from your Oxford Nanopore sequencing run
• Reference tRNA sequences with adapters (or use the provided S. cerevisiae reference)

Step 1: Organize Your Data¶

The pipeline expects POD5 files organized by sequencing run:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

/path/to/your/data/
├── run1/
│   ├── pod5_pass/     # Passed reads (required)
│   │   ├── file1.pod5
│   │   └── file2.pod5
│   └── pod5_fail/     # Failed reads (optional)
│       └── file3.pod5
└── run2/
    └── pod5/          # Alternative directory name
        └── file4.pod5

The pipeline searches for POD5 files in pod5_pass/, pod5_fail/, or pod5/ subdirectories.

Step 2: Create Sample File¶

Create a tab-separated sample file listing your samples:

1
2
3

sample1 /path/to/your/data/run1
sample2 /path/to/your/data/run2
sample1 /path/to/your/data/run1_replicate

For multiplexed samples with barcodes, use YAML format instead. See Sample Files.

Step 3: Create Configuration¶

Create a configuration file for your analysis:

1
2
3
4
5
6
7
8
9

# Sample file path
samples: config/samples-myproject.tsv

# Output directory
output_directory: "results/myproject"

# Optional: Override base configuration
# fasta: "path/to/custom/reference.fa"
# remora_cca_classifier: "path/to/custom/model.pt"

The configuration inherits defaults from config/config-base.yml. Override any parameter as needed.

Step 4: Dry Run¶

Preview the pipeline execution:

1

pixi run snakemake -n --configfile=config/config-myproject.yml

Verify:

• All samples are detected
• Expected number of jobs
• No errors in DAG construction

Step 5: Run Pipeline¶

Local Execution¶

For small datasets or testing:

1

pixi run snakemake --cores 12 --configfile=config/config-myproject.yml

Adjust --cores based on your system.

Cluster Execution (Recommended)¶

For production runs, use cluster execution:

1

pixi run snakemake --profile cluster/lsf --configfile=config/config-myproject.yml

1

pixi run snakemake --profile cluster/slurm --configfile=config/config-myproject.yml

See Cluster Setup for detailed cluster configuration.

Background Execution¶

For long-running jobs, run in background:

1
2
3

nohup pixi run snakemake --profile cluster/lsf \
    --configfile=config/config-myproject.yml \
    > pipeline.log 2>&1 &

Monitor progress:

1

tail -f pipeline.log

Step 6: Monitor Progress¶

Check Job Status¶

1
2
3
4
5

# For LSF
bjobs -u $USER

# For SLURM
squeue -u $USER

View Snakemake Progress¶

1
2
3
4

# Show running jobs
pixi run snakemake --profile cluster/lsf \
    --configfile=config/config-myproject.yml \
    --summary

Check Logs¶

Rule-specific logs are in the output directory:

1

ls results/myproject/logs/

Step 7: Verify Outputs¶

After completion, check your outputs:

1

ls -la results/myproject/summary/tables/

Key Files to Check¶

1. Charging CPM - Per-tRNA charging quantification:

   Bash
   1	zcat results/myproject/summary/tables/sample1/sample1.charging.cpm.tsv.gz | column -t | head

2. Alignment Stats - Read counts through pipeline:

   Bash
   1	zcat results/myproject/summary/tables/sample1/sample1.align_stats.tsv.gz | column -t

3. Final BAM - Verify charging tags:

   Bash
   1	samtools view results/myproject/bam/final/sample1/sample1.bam | head -1

Troubleshooting¶

No POD5 Files Found¶

1

Error: No POD5 files found for sample 'sample1'

Solution: Verify POD5 files exist in pod5_pass/, pod5_fail/, or pod5/ subdirectories.

GPU Memory Error¶

1

CUDA out of memory

Solution: Reduce batch size or run one GPU job at a time. Adjust cluster profile ngpu resource.

Rule Failed¶

1

Error in rule rebasecall

Solution: Check rule-specific log:

1

cat results/myproject/logs/rebasecall/sample1.log

See Troubleshooting for more solutions.

Next Steps¶

• Output Files - Understand all output files
• Configuration - Customize parameters
• Workflow Overview - Understand the pipeline stages