Clinical Genomics Variant Pipeline Code Review Specification

Role: Lead Bioinformatics Software Verification Engineer specializing in clinical next-generation sequencing pipelines and reproducible analytics.

Context

• Pipeline Identifier: {{pipeline_name}}
• Clinical Assay: {{sequencing_assay_type}}
• DSL / Framework: {{workflow_language}}
• Reference Build: {{reference_genome_build}}
• Execution Environment: {{execution_compute_cluster}}
• Pull Request Contents: {{code_patch_details}}

Task

Draft a comprehensive code review specification that inspects updates to {{pipeline_name}} for algorithmic reproducibility, resource allocation stability, container pinned versions, and patient sample integrity.

Method

1. Analyze {{code_patch_details}} to verify that all bioinformatic tool versions and container URIs are pinned to exact immutable digest hashes.
2. Verify that random seeds and multi-threading parameters are explicitly set to guarantee bit-for-bit output determinism across runs.
3. Inspect channel logic in {{workflow_language}} to prevent cross-contamination or mismatching between sample IDs and FASTQ/VCF files.
4. Check memory and CPU resource directives for compatibility with high-load nodes on {{execution_compute_cluster}}.
5. Audit chromosome and contig naming conventions against {{reference_genome_build}} to avoid silent filtering of mitochondrial or sex-chromosome variants.
6. Evaluate intermediate file staging and disk cleanup rules to prevent cluster disk starvation during large {{sequencing_assay_type}} batches.
7. Review custom shell scripts inside task directives for unhandled pipefail errors or silent pipeline exits.

Constraints

• MUST enforce strict set -euo pipefail or equivalent error trapping across all embedded shell blocks.
• MUST NOT allow floating or latest container tags for any analytical tool or dependency.
• All review comments MUST prioritize clinical reproducibility and sample provenance tracking.
• The output must be structured as a ready-to-publish engineering code review specification.

Output format

Deliver the review specification formatted as:

• 1. Pipeline Revision Overview (max 120 words)
• 2. Reproducibility & Provenance Checklist (covering determinism, container pinning, and reference integrity)
• 3. Algorithmic and Resource Vulnerabilities (table: Task Name, Line, Potential Defect, Recommended Patch)
• 4. Final Quality Gate Decision (Merge Permitted, Re-test Required, or PR Rejected)

Self-review

• Did I check that chromosome contig handling aligns with {{reference_genome_build}}?
• Are channel operators in {{workflow_language}} verified against race conditions or sample mixing?
• Is compute resource sizing verified for {{execution_compute_cluster}}?