usage_examples

Usage Examples

Practical examples for using SQANTI-browser.

---

🚀 Quick Start

Most Basic Command

python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --output my_hub \
    --genome hg38

That's it! This creates a track hub in my_hub/ ready to upload.

Recommended Command (With Tables)

python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --output my_hub \
    --genome hg38 \
    --tables

Adds interactive HTML reports for exploring your data offline.

---

📦 After Generation → Upload

See Hosting Guide for detailed upload instructions:

• GitHub (easiest)
• Institutional server
• Cloud storage (AWS S3, Google Cloud)

Quick summary: Upload files → Get public URL to hub.txt → Load in UCSC (My Data → Track Hubs)

---

🎯 Common Use Cases

1. Add All Validation Tracks (Recommended!)

python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --genome hg38 \
    --output my_hub \
    --refGTF reference.gtf \
    --star-sj SJ.out.tab \
    --CAGE-peak CAGE_peaks.bed \
    --polyA-peak polyA_peaks.bed \
    --tables

What this does:

• Shows your transcripts alongside reference annotation
• Adds short-read junction support
• Validates transcription start sites (CAGE)
• Validates termination sites (polyA)
• Creates interactive HTML tables

When to use: Publication-quality visualization with full validation

---

2. Custom Genome (Non-Model Organism)

python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --genome my_species_v1 \
    --twobit genome.2bit \
    --output my_hub

When to use: Working with species not in UCSC (e.g., plants, non-model animals)

Note: --twobit automatically extracts chromosome sizes. See Custom Genomes for details.

---

3. Large Dataset (Faster Processing)

python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --genome hg38 \
    --output my_hub \
    --no-category-tracks

When to use: >50K transcripts, want faster processing

Trade-off: Only creates main track (all transcripts), skips individual category tracks

---

4. Subset of Category Tracks

python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --genome hg38 \
    --output my_hub \
    --category-tracks FSM,ISM,NIC

When to use: Want only FSM, ISM, and NIC tracks (not all nine categories)

Valid abbreviations: FSM, ISM, NIC, NNC, antisense, genic_intron, genic_genomic (or genic), intergenic, fusion

---

5. Sort by Expression (Show Top Expressed First)

python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --genome hg38 \
    --output my_hub \
    --sort-by FL

When to use: Want to prioritize viewing highly expressed isoforms

Other sort options: length, FL, diff_to_TSS, diff_to_TTS, dist_to_CAGE_peak, dist_to_polyA_site

See Isoform Ordering for details.

---

🔬 Validation Tracks Deep Dive

STAR Junctions (Short-Read Support)

--star-sj path/to/SJ.out.tab

What it does: Shows which junctions have short-read support from STAR alignment

When to use: You have paired short+long read data

---

CAGE Peaks (TSS Validation)

--CAGE-peak path/to/CAGE_peaks.bed

What it does: Validates transcription start sites (5' ends)

Where to get: See SQANTI3 CAGE documentation

When to use: Assess 5' end accuracy

---

PolyA Peaks (TTS Validation)

--polyA-peak path/to/polyA_peaks.bed

What it does: Validates transcription termination sites (3' ends)

Where to get: See SQANTI3 polyA documentation

When to use: Assess 3' end accuracy

---

Reference Annotation

--refGTF path/to/reference.gtf

What it does: Shows reference annotation alongside your transcripts for comparison

Where to get:

• Same file you used with SQANTI3
• GENCODE (human/mouse)
• Ensembl (all species)
• SQANTI3 example (chr22 test data)

When to use: Almost always! Shows context for novel isoforms

---

🛠️ Testing & Debugging

Command	What It Does	When to Use
--validate-only	Check inputs, don't create hub	Test before full run
--dry-run	Process data, skip bigBed creation	Test logic/filters
--keep-temp	Keep temporary files	Debug issues

Example:

# Quick validation (1 second)
python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --genome hg38 \
    --output my_hub \
    --validate-only

# Full dry run (processes data but doesn't create bigBeds)
python -m sqanti_browser \
    --gtf corrected.gtf \
    --classification classification.txt \
    --genome hg38 \
    --output my_hub \
    --dry-run

---

📊 Standalone HTML Reports

Generate reports without creating a track hub:

python src/filter_isoforms.py \
    --classification classification.txt \
    --output-dir reports

With ORF sequences:

python src/filter_isoforms.py \
    --classification classification.txt \
    --output-dir reports \
    --include-sequences

When to use: Share data tables with collaborators who don't need UCSC visualization

---

🎓 Example Dataset

Run the example workflow script to see commands and parameters:

python example/example_usage.py

Try SQANTI-browser with the included example:

# Basic example
python -m sqanti_browser \
    --gtf example/SQANTI3_QC_output/example_corrected.gtf \
    --classification example/SQANTI3_QC_output/example_classification.txt \
    --output example_output \
    --genome hg38 \
    --tables

# With all validation tracks
python -m sqanti_browser \
    --gtf example/SQANTI3_QC_output/example_corrected.gtf \
    --classification example/SQANTI3_QC_output/example_classification.txt \
    --genome hg38 \
    --output example_output \
    --chrom-sizes example/SQANTI3_QC_output/chrom.sizes \
    --star-sj example/SQANTI3_QC_output/exampleSJ.out.tab \
    --CAGE-peak example/SQANTI3_QC_output/chr22.human.refTSS_v3.1.hg38.bed \
    --polyA-peak example/SQANTI3_QC_output/polyApeaks.atlas.GRCh38.bed \
    --tables

---

💡 Quick Tips

Tip	Why
Always use --tables	Interactive reports are super useful
Add --refGTF	Compare with reference annotation
Use --validate-only first	Catch errors before full run
Run hubCheck hub.txt	Validate before upload
Sort by FL	See highest expressed isoforms first
Use --no_highlight	Disable highlighted isoforms

---

📚 See Also

• Quick Reference - One-page cheat sheet
• Command Line Reference - All options explained
• Hosting Guide - Upload your hub
• Non-reference Genomes - Working with .2bit files
• SQANTI-reads Integration - Multi-sample experiments

---