Interview-Driven Workflow Construction and Repair in Galaxy

Galaxy Workflow Data-Flow Brief: Ebola Virus Infection RNA-seq Analysis

Data-Flow Overview

The workflow transforms raw paired-end RNA-seq collections into sample-level transcript abundance estimates and downstream gene-level differential expression statistics.

[reads_paired_collection (list:paired)] ──┐
                                          ├─► [Step 1: salmon_quant (map-over)] ──► [salmon_quant_collection (list)]
[salmon_index] ───────────────────────────┘                                              │
                                                                                         ▼
[sample_metadata (tabular)] ───────────────────────────────────────────────────────────► [Step 2: deseq2 (reduce)] ──► [deseq2_norm_counts]
                                                                                         ▲                           ──► [deseq2_differential_results]
[tx2gene_map (tabular)] ─────────────────────────────────────────────────────────────────┘                           ──► [deseq2_plots]

Step Details and Collection Operations

Step 1: salmon_quant (Map-over)

• Operation: Pseudoalignment and transcript abundance quantification.
• Input Wiring:
  • paired_reads: Connected to reads_paired_collection (mapped over each paired element).
  • index: Connected to salmon_index.
• Collection Behavior: Maps over the input list:paired collection, preserving sample element identifiers.
• Output: quant.sf transcript quantification table per element, producing a list collection salmon_quant_collection.

Step 2: deseq2 (Reduce / Collective modeling)

• Operation: Multi-factor differential gene expression modeling using DESeq2.
• Input Wiring:
  • counts_or_abundance: Connected to salmon_quant_collection (accepting the collection of quantification tables).
  • sample_table: Connected to sample_metadata.
  • tx2gene: Connected to tx2gene_map.
• Collection Behavior: Reduces the collection of sample abundances into collective matrix modeling.
• Outputs:
  • normalized_counts: Tabular normalized expression values across samples.
  • differential_expression: Tabular results with log2 fold-changes, p-values, and adjusted FDR.
  • qc_plots: Diagnostic PDF containing PCA and dispersion plots.

Galaxy Tool Equivalents & Unresolved Needs

• salmon_quant: Standard Galaxy tool toolshed.g2.bx.psu.edu/repos/bgruening/salmon/salmon/1.10.0+galaxy0 (or compatible version).
• deseq2: Standard Galaxy tool toolshed.g2.bx.psu.edu/repos/iuc/deseq2/deseq2/2.11.40.8+galaxy0 (or compatible version).
• No custom wrappers required; standard ToolShed tools fulfill all operations.

Confidence and Open Questions

• High confidence in map-over on list:paired feeding directly into DESeq2’s collection-aware input interface or standard tabular aggregation.
• Carried open requirement: Verify DESeq2 parameter interface for tximport/tx2gene integration vs feeding gene-level count tables directly.