Interview-Driven Workflow Construction and Repair in Galaxy

Galaxy Workflow Interface Brief: Ebola Virus Infection RNA-seq Analysis

Workflow Purpose & Overview

Process paired-end RNA-seq reads from human A549 WT and A549 PTPN13-KO cells infected with Ebola virus or mock at 24 and 48 HPI. The workflow performs pseudoalignment and transcript abundance quantification with Salmon, aggregates abundances to gene-level counts, and conducts multi-factor differential expression analysis with DESeq2.

Workflow Inputs

Input Identifier	Label	Type	Format / Collection Shape	Optional / Default	Description
reads_paired_collection	Paired-end RNA-seq reads	data_collection_input	list:paired (fastqsanger, fastqsanger.gz)	Required	Dataset collection of paired-end Illumina FASTQ reads across experimental samples
salmon_index	Salmon reference transcriptome index	data_input	tar, directory (or fasta)	Required	Precomputed Salmon index directory/archive for human transcriptome targets
sample_metadata	Sample metadata / Experimental factors	data_input	tabular	Required	Tabular design matrix specifying sample IDs, Genotype (WT / KO), Condition (Mock / Ebola), and Timepoint (24 / 48)
tx2gene_map	Transcript-to-gene mapping table	data_input	tabular	Required	Two-column tab-delimited mapping of transcript IDs to gene IDs

Workflow Outputs

Output Identifier	Label	Source Step / Producer	Datatype / Collection	Public / Checkpoint	Description
salmon_quant_collection	Salmon quantification tables (quant.sf)	Salmon step	list of tabular	Checkpoint / Public	Per-sample transcript abundance estimates and counts
deseq2_norm_counts	DESeq2 normalized counts matrix	DESeq2 step	tabular	Public	Size-factor normalized gene expression matrix across all samples
deseq2_differential_results	DESeq2 differential expression results	DESeq2 step	tabular	Public	Statistical test statistics (log2FC, p-value, padj/FDR) for tested contrasts
deseq2_plots	DESeq2 diagnostic plots (PCA & dispersions)	DESeq2 step	pdf	Public	Quality control PCA and dispersion diagnostic plots

Collection Shapes & Mapping Strategy

• Reads Collection: Structured as list:paired containing each biological sample as an element mapped to forward and reverse FASTQ datasets.
• Salmon Execution: Mapped over the list:paired collection, outputting a list of quant.sf tabular datasets.
• Aggregation & Modeling: DESeq2 takes the collection of count/abundance tables together with sample_metadata and tx2gene_map as collective inputs to model variance and differential expression across the 8 experimental groups (2 genotypes × 2 conditions × 2 timepoints).

Confidence and Assumptions

• High confidence in list:paired collection shape for paired-end Illumina RNA-seq.
• High confidence in Salmon + DESeq2 toolchain as standard, well-supported Galaxy ecosystem tools.
• Checkpoint output salmon_quant_collection allows isolated verification of the quantification tier before differential modeling.

Open Questions & Interface Obligations

1. Whether Salmon index is provided as a pre-built index directory/tarball or built dynamically from a FASTA file.
2. The specific DESeq2 design formula (additive ~ Genotype + Time + Condition vs full interaction ~ Genotype * Time * Condition).