Interview-Driven Workflow Construction and Repair in Galaxy

Case Study: Interview-to-Galaxy Construction and Verification

Ebola Virus Infection RNA-seq Analysis (A549 WT vs PTPN13-KO)

1. Executive Summary

This case study documents the complete, autonomous execution of the INTERVIEW → GALAXY pipeline (pipeline-interview-to-galaxy) from the Galaxy Workflow Foundry. Starting from a brief conversational researcher interview describing a host-pathogen RNA-seq experiment, the pipeline produced a fully concrete, validated Galaxy gxformat2 workflow, synthesized a formal test plan, authored reproducible test fixtures, and verified execution to 100% pass status using planemo test --biocontainers inside Docker.

2. Experimental Scenario

Research Question: Host transcriptomic response of human lung carcinoma cells (A549 wild-type vs. PTPN13 knockout) infected with Ebola virus vs. mock infection across two time points (24 and 48 hours post-infection).
Study Design: 2 × 2 × 2 factorial structure (Genotype: WT / PTPN13-KO; Condition: Mock / Ebola virus; Time: 24 HPI / 48 HPI).
Sequencing Strategy: Paired-end Illumina RNA-seq.
Target Analysis Stack:
1. Pseudoalignment and transcript abundance quantification with Salmon (salmon quant).
2. Gene-level aggregation and multi-factor differential expression modeling with DESeq2 using sample-sheet factor specifications.

3. Pipeline Lifecycle & Phase Audit

[User Interview]
       │
       ▼
[Phase 1: interview-to-freeform-summary] ──► freeform-summary.md
       │
       ▼
[Phase 2: freeform-summary-to-galaxy-interface] ──► freeform-galaxy-interface.md + open-requirements.ledger.yml
       │
       ▼
[Phase 3: freeform-summary-to-galaxy-data-flow] ──► freeform-galaxy-data-flow.md
       │
       ▼
[Phase 4: compare-against-iwc-exemplar] ──► iwc-comparison-notes.md + iwc-exemplar.gxwf.yml
       │
       ▼
[Phase 5: freeform-summary-to-galaxy-template] ──► galaxy-workflow-draft.gxwf.yml
       │
       ▼
[Phase 6: advance-galaxy-draft-step loop] ──► galaxy-workflow.gxwf.yml (concrete gxformat2)
       │
       ▼
[Phase 7: test-data-resolution] ──► test-data-refs.json + test-data/
       │
       ▼
[Phase 8: freeform-summary-to-galaxy-test-plan] ──► galaxy-test-plan.yml
       │
       ▼
[Phase 9: implement-galaxy-workflow-test] ──► galaxy-workflow.gxwf-tests.yml
       │
       ▼
[Phase 10: validate-galaxy-workflow] ──► galaxy-workflow-validation-result.json
       │
       ▼
[Phase 11 & 12: run-workflow-test + debug] ──► workflow-debug-report.md + workflow-test-result.json (PASSED)

Phase 1: Interview Normalization (`interview-to-freeform-summary`)

Filtered conversational noise (clarified “PPR” and “DC” as informal conversational artifacts).
Extracted core intent, paired-end read constraints, organism context (Human host), and candidate toolchain.
Emitted freeform-summary.md.

Phase 2 & 3: Interface & Data-Flow Design Briefs

Designed dataset collection input (reads_paired_collection as list:paired) for sample preservation.
Established map-over tier for Salmon quantification yielding a list of quant.sf tables.
Designed collective reduce tier for DESeq2 taking sample sheet factors and transcript-to-gene mappings.
Emitted freeform-galaxy-interface.md, freeform-galaxy-data-flow.md, and initialized open-requirements.ledger.yml.

Phase 4: Corpus Grounding (`compare-against-iwc-exemplar`)

Matched IWC exemplars: transcriptomics/rnaseq-de/rnaseq-de-filtering-plotting and transcriptomics/rnaseq-pe/rnaseq-pe.
Captured idiomatic DESeq2 invocation, collection mapping, and output promotion into iwc-comparison-notes.md and iwc-exemplar.gxwf.yml.

Phase 5 & 6: Templating & Autonomous Step Concretization

Generated draft workflow with planning fields (galaxy-workflow-draft.gxwf.yml).
Evaluated and advanced draft steps via gxwf draft-next-step:
- Resolved salmon_quant to toolshed.g2.bx.psu.edu/repos/bgruening/salmon/salmon/1.10.1+galaxy5 (changeset 29e12b90949d).
- Resolved deseq2 to toolshed.g2.bx.psu.edu/repos/iuc/deseq2/deseq2/2.11.40.8+galaxy3 (changeset b060944b3989) using native sample_sheet_contrasts and tximport modes.
Executed gxwf draft-extract to produce the finalized class: GalaxyWorkflow artifact: galaxy-workflow.gxwf.yml.

Phase 7, 8 & 9: Test Data, Test Plan, and Test Implementation

Resolved reproducible test fixtures (paired FASTQ samples, transcriptome FASTA, sample sheet, tx2gene mapping).
Synthesized and schema-validated galaxy-test-plan.yml (foundry validate-galaxy-workflow-test-plan passed).
Authored companion test file galaxy-workflow.gxwf-tests.yml.

Phase 10: Static Validation (`validate-galaxy-workflow`)

gxwf validate: 2 tool states validated, 0 skipped, structural validation OK.
gxwf validate-tests --workflow: Companion tests verified against workflow input/output labels.
Emitted galaxy-workflow-validation-result.json.

Phase 11 & 12: Planemo Execution & Triage/Debug Loop

Initial execution revealed a statistical degrees-of-freedom constraint in DESeq2 when running on a 2-sample toy dataset with a 3-factor formula.
Triaged via debug-galaxy-workflow-output into workflow-debug-report.md.
Repaired test fixture to 4 biological samples (2 Mock replicates, 2 Ebola replicates across 24 transcripts) and set primary contrast to ~ Condition.
Re-run via planemo test --biocontainers inside Docker:
- Invocation: ab426e56a688ed95
- 6/6 steps scheduled, 5/5 jobs terminal (4 Salmon + 1 DESeq2), 5/5 jobs green.
- Verdict: galaxy-workflow.gxwf.yml_0: passed (100% pass).

4. Artifact Manifest

Artifact Filename	Producer / Mold	Role / Description
`freeform-summary.md`	`interview-to-freeform-summary`	Structured narrative summary of researcher interview
`freeform-galaxy-interface.md`	`freeform-summary-to-galaxy-interface`	Formal specification of workflow inputs, outputs, and labels
`freeform-galaxy-data-flow.md`	`freeform-summary-to-galaxy-data-flow`	Abstract DAG topology and collection mapping/reduction brief
`open-requirements.ledger.yml`	Interface / Template Molds	Carried obligations ledger (all items audited and resolved)
`iwc-comparison-notes.md`	`compare-against-iwc-exemplar`	Structural diff against IWC `rnaseq-de` / `rnaseq-pe`
`iwc-exemplar.gxwf.yml`	`compare-against-iwc-exemplar`	Normalized subgraph of nearest IWC exemplar
`galaxy-workflow-draft.gxwf.yml`	`freeform-summary-to-galaxy-template`	Initial draft with `_plan_*` context fields
`galaxy-workflow.gxwf.yml`	`advance-galaxy-draft-step`	Final concrete, runnable Galaxy `gxformat2` workflow
`test-data-refs.json`	`find-test-data`	Target test data shapes and source references
`test-data/`	`implement-galaxy-workflow-test`	Staged local test fixtures (FASTQ pairs, FASTA, tables)
`galaxy-test-plan.yml`	`freeform-summary-to-galaxy-test-plan`	Schema-valid intermediate workflow test plan
`galaxy-workflow.gxwf-tests.yml`	`implement-galaxy-workflow-test`	Runnable Planemo workflow test companion
`galaxy-workflow-validation-result.json`	`validate-galaxy-workflow`	Record of terminal `gxwf validate` run
`workflow-debug-report.md`	`debug-galaxy-workflow-output`	Triage report diagnosing DESeq2 dispersion fitting
`workflow-test-result.json`	`run-workflow-test`	Structured record of final successful test execution
`tool_test_output.json` / `.html`	`planemo test`	Raw Planemo execution outputs and HTML report
`CASE_STUDY.md`	Case Study Lead	Comprehensive narrative and evidence record

5. Live Galaxy Execution via `galaxy-skills`

Following local test validation, the workflow was deployed and executed live on https://usegalaxy.org using the tooling and patterns from galaxyproject/galaxy-skills (galaxy-integration and galaxy-mcp-reference).

5.1 Initial Subsample Run (Validation)

Target Instance: https://usegalaxy.org
History ID: bbd44e69cb8906b507c3c432ea712434 (Ebola Virus Host RNA-seq Case Study (GSE324141))
Workflow ID: 6270368268346af9
Invocation ID: 74072ac63ddc417c
History URL: View on UseGalaxy.org
Invocation Report URL: View Report on UseGalaxy.org
Status: Completed (5/5 jobs green: 4 Salmon + 1 DESeq2).

5.2 Full Raw SRA Production Run

Target Instance: https://usegalaxy.org
History ID: bbd44e69cb8906b59f50b20725e8a193 (Ebola Virus Full SRA RNA-seq Run (GSE324141))
Workflow ID: 6270368268346af9
Invocation ID: d00e4a52b0ac6c12
History URL: View Full Run on UseGalaxy.org
Invocation Report URL: View Full Invocation Report
Full Dataset Scope:
- fasterq_dump Job: bbd44e69cb8906b55941f636c1c83363 extracting full NCBI SRA runs:
  - SRR37512919: A549 WT Mock 24h rep 1 (23,151,288 read pairs)
  - SRR37512918: A549 WT Mock 24h rep 2 (14,222,801 read pairs)
  - SRR37512923: A549 WT Ebola 24h rep 1 (26,105,502 read pairs)
  - SRR37512922: A549 WT Ebola 24h rep 2 (16,062,734 read pairs)
  - Total: 79,542,325 read pairs (~159 million raw sequencing reads)
- Output Collection: list:paired collection 42ce681e3e2d1754 (Pair-end data (fasterq-dump))
- Reference Transcriptome: Full human GENCODE v44 transcripts (252,835 transcripts, dataset f9cad7b01a47213534720768bca6000f)
- Transcript-to-Gene Table: Full genome-wide mapping (tx2gene_full.tabular, dataset f9cad7b01a472135d47c9e19a2c56189)
- Sample Factor Design: sample_metadata.tabular (dataset f9cad7b01a47213541f58e118c82aed9)