How nf-core/rnaseq works
Main Components
• Input Data and Preprocessing:
– Show the Samplesheet and FASTQ files as the initial data inputs.
– Include preprocessing steps such as merging FASTQ files (using “cat”), quality control (e.g. FastQC), UMI extraction, adapter/quality trimming (Trim Galore!), and removal of contaminants.
• Alignment/Quantification Modules:
– Display a split into multiple quantification paths: for example, one branch using STAR alignment followed by Salmon quantification; an alternative STAR → RSEM path; and a branch for HISAT2 (which in this case does not perform quantification).
– Add nodes for tools like STAR, HISAT2, Salmon, and RSEM.
• Data Processing and Post-Alignment:
– Include nodes for alignment post-processing using tools such as SAMtools (for sorting and indexing), duplicate marking (Picard MarkDuplicates and UMI-tools deduplication), and transcript assembly (StringTie) as needed.
• Quality Control and Reporting:
– Represent the comprehensive QC components including tools like RSeQC, Qualimap, dupRadar, Preseq, and the final MultiQC aggregation.
• Configuration and Modular Setup:
– Illustrate the “main.nf” as the central Nextflow entry point.
– Indicate the “modules” and “subworkflows” directories which encapsulate individual functions (with local modules and nf-core modules distinguished).
– Show configuration (conf/* and nextflow.config) which govern pipeline behavior, profiles (docker, singularity, conda), and test environments.
• Deployment & Environment:
– Represent containerization options (Docker, Singularity) and the support for cloud environments (as indicated by the Seqera Platform and GitHub Actions workflows).
Relationships and Data Flow
• How alignment outputs flow into post-alignment processing (sorting, indexing, deduplication) and then into the QC modules.
Architectural Patterns or Design Principles
• Modular and Configurable: Emphasize that the system is built using modular Nextflow components (modules, subworkflows) which can be independently developed and tested.
• Pipeline-as-Code: Indicate the declarative nature (DSL2) where pipeline steps, dependencies, and parameters are defined in code.
• Containerization & Portability: Show that container images (Docker, Singularity) are part of the design to ensure reproducibility.
• Continuous Integration: Reference the presence of CI workflows (under .github/workflows) that automate linting, testing, and deployment.
Relevant Technologies and Frameworks
• Nextflow DSL2 (workflow orchestration)
• nf-core modules and community guidelines
• Container technologies: Docker, Singularity, Conda for dependency management
• Bioinformatics tools (FastQC, Trim Galore!, STAR, HISAT2, Salmon, RSEM, SAMtools, Picard, MultiQC, etc.)
Additional Details
• Structure emphasizes the modular, extensible nature of the pipeline, capturing each of the functional components described in the README and reflected in the file tree.
• Integration points for external dependencies and document tests and CI/CD as auxiliary infrastructure ensuring code quality.
Happy diagramming!