Skip to content

AGENTS.md

Latest commit

 

History

History
106 lines (80 loc) · 2.91 KB

AGENTS.md

File metadata and controls

106 lines (80 loc) · 2.91 KB

Agent Directives: config/amalgkit

Role

Production-ready amalgkit RNA-seq workflow configurations for automated transcript quantification pipelines.

Contents

File Description
amalgkit_template.yaml Reference: 400+ line template with all options documented
amalgkit_test.yaml Minimal test configuration for validation
amalgkit_pogonomyrmex_barbatus.yaml Production: Full P. barbatus dataset (95/110 quantified)
amalgkit_apis_mellifera_all.yaml Production: Full A. mellifera dataset (~7,270 samples)
amalgkit_cross_species.yaml Cross-species TMM normalization config
tissue_mapping.yaml Canonical tissue name synonyms for normalization
tissue_patches.yaml Per-bioproject/sample tissue overrides

Configuration Structure

# Core paths (relative to repo root)
work_dir: output/amalgkit/{species}/work
log_dir: output/amalgkit/{species}/logs
threads: 16

# Species identification
species_list:
  - Pogonomyrmex_barbatus
taxon_id: 144034

# Reference genome
genome:
  accession: GCF_000187915.1
  dest_dir: output/amalgkit/shared/genome/Pogonomyrmex_barbatus

# Step-specific parameters
steps:
  getfastq:
    redo: no           # Skip already-downloaded
    keep_fastq: no     # Delete after quant
  quant:
    redo: no           # Skip already-quantified
    index_dir: ...     # Reuse kallisto index

Critical Patterns

Stream-and-Clean (Disk Management)

For large datasets with limited disk space:

steps:
  getfastq:
    redo: no           # Resume capability
  quant:
    keep_fastq: no     # Immediate cleanup
    redo: no           # Idempotent

Shared Resources

Reuse genome/index across configs:

genome:
  dest_dir: output/amalgkit/shared/genome/Pogonomyrmex_barbatus
steps:
  quant:
    index_dir: output/amalgkit/shared/genome/Pogonomyrmex_barbatus/index

Metadata Filtering

Filter to RNA-Seq + Illumina to prevent genomic samples leaking in:

steps:
  metadata:
    search_string: '"Species"[Organism] AND "RNA-Seq"[Strategy] AND "Illumina"[Platform]'

Adding New Species

  1. Copy amalgkit_template.yaml → amalgkit_{species}.yaml
  2. Update species_list, taxon_id, and genome.accession
  3. Adjust paths: work_dir, log_dir, genome.dest_dir
  4. Validation: Run python3 scripts/rna/validate_configs.py to ensure schema compliance.
  5. Test with small sample subset first (use max_sample: 5)
  6. Scale to full dataset after validation

Validation and Testing

  • Config Validation: usage of scripts/rna/validate_configs.py is mandatory for all new configurations.
  • Zero-Mock Policy: All Amalgkit tests strictly adhere to the Zero-Mock policy, ensuring real functional verification of the CLI and environment.

Environment Overrides

Prefix with AK_:

export AK_THREADS=16
export AK_WORK_DIR=/fast/storage/amalgkit
export NCBI_EMAIL=your@email.com