How oskarvid/selma works
Main Components
a. Input Layer:
• FASTQ files input directory
• Reference genome files
• Configuration files (config.yaml, samples.tsv)
b. Workflow Management Layer:
• Snakemake workflow engine
• Environment management systems (Conda/Docker/Guix)
c. Processing Pipeline Components:
• Alignment Module (BWA)
• BAM Processing Module (Samtools)
• Quality Control Module (GATK4 tools)
• Variant Calling Module (GATK4 HaplotypeCaller)
• Variant Processing Module (GATK4 GenotypeGVCFs, VariantRecalibrator)
[Input Layer]
• ↓
[Environment Management]
• ↓
[Workflow Engine]
• ↓
[Processing Pipeline]
Specific Elements
a. Input Section:
b. Environment Section:
• Docker container
• Conda environment
• Guix shell
c. Workflow Engine Section:
• Snakemake as the central orchestrator
• Scatter-gather parallelization mechanism
d. Processing Pipeline Section:
• Create sequential blocks for each major processing step:
1. Read Mapping (BWA)
2. BAM Processing (Samtools)
3. Duplicate Marking (GATK4)
4. Base Recalibration (GATK4)
5. Variant Calling (HaplotypeCaller)
6. Variant Genotyping (GenotypeGVCFs)
7. Variant Recalibration (VQSR)
Important Details to Highlight
• Blue: Input/Output data
• Green: Processing tools
• Yellow: Configuration and environment
• Red: Critical processing steps
• Parallel processing capabilities:
• Where scatter-gather operations occur
• Thread distribution
• Resource requirements:
• Memory requirements (64GB RAM recommended)
• Storage requirements (500GB minimum)
• CPU requirements (16 threads optimal)
Additional Elements
• Version numbers for key tools
• Temporary file storage locations
• Optimization points for resource usage
Performance Characteristics
A small section showing:
• Typical runtime (16-18 hours for 30x coverage)
• Resource scaling information
• Bottleneck points
This system design should effectively communicate both the high-level architecture and the detailed workflow of the Selma pipeline, making it valuable for both new team members and experienced developers.