Dataset DOI: 10.5061/dryad.g1jwstr5k

Description of the data and file structure

This dataset contains bulk RNA-seq processing and downstream analyses examining sources of transcriptomic variation (batch, colony, patriline, caste, and tissue) in Vespula wasps. The pipeline integrates differential gene expression (DESeq2), alternative splicing (rMATS), weighted gene co-expression network analysis (WGCNA), tissue-specificity indices (tau), functional annotation (eggNOG-mapper), GO term enrichment (topGO) with semantic-similarity reduction (rrvgo), pairwise/set-level comparisons, correlation analyses, and Cytoscape network exports.

Files are grouped into numbered top-level archives (*.tar.gz) reflecting the sequential steps of the analysis. Each archive expands to a directory of the same name. Associated scripts can be found in zipped (*.zip) files of the same name.

Files and variables

Top-Level Archive Contents

---

Factor / Sample Naming Conventions

Filenames throughout use a consistent nomenclature. Understanding these tokens will help interpret all files:

• Factors: Batch, Caste, Tissue, Patriline, Colony — the design variable being contrasted.
• Caste levels: Queen, Worker
• Tissue levels: Head, Thorax
• Colony IDs: numeric (e.g., 232, 247, 253, 254, 255, 256, 258)
• Patriline IDs: <Colony>_P<n> (e.g., 232_P4, 256_P1) — indicates the patriline within a colony
• Batch IDs: numeric (e.g., 5908, 6011-Test, 6136, 6137, 6269, 6270) — sequencing batches
• Comparison direction: <group1>_vs_<group2> (DESeq2 style, with underscores) or <group1>.vs.<group2> (rMATS style, with dots). Both directions are typically included so that up-/down-regulated genes can be extracted from either perspective.
• GO ontology suffixes: BP = Biological Process, MF = Molecular Function, CC = Cellular Component.
• rMATS event types: SE (skipped exon), A5SS (alternative 5′ splice site), A3SS(alternative 3′ splice site), MXE (mutually exclusive exons), RI (retained intron).
• rMATS read-count schemes: JC = junction counts only; JCEC = junction counts + exon-body counts.

---

0_datasets/

Raw/processed inputs to the pipeline: sample metadata (sample IDs, colony, patriline, caste, tissue, batch assignments), gene-level count matrices, and reference annotation files. This is the starting point for all downstream steps.

Primary reference files

vespula_samples.csv

253 rows, one per sample. Columns: Record_Name (short internal sample code), Batch, Colony, Patriline, ID (individual number), Caste, Tissue, Complete_Name (the full identifier used in downstream file names

vespula_traits.csv

253 rows, one per sample; 33 columns. An encoded design matrix. Each row is a sample (indexed by Complete_Name); each column after the first is a batch, colony, patriline, caste, or tissue level, with 1 indicating membership and 0 otherwise. This matrix underlies the model/design formulas used in WGCNA.

---

7_deseq2/ — Differential Gene Expression

Top-level files

DESeq2 results structured by the model design used:

• colony/ — Models where colony is included as a factor (colony-level analysis, using pooled samples per colony).
• patriline/ — Models using individual patriline-level replicates (finer-grained within-colony analysis).

Within each design, filenames correspond to the contrast factor: Batch, Caste, Tissue, Colony, or Patriline. Files typically include:

• Per-contrast results tables (.csv) with the standard DESeq2 columns:
  • gene_id (or rowname) — gene identifier
  • baseMean — mean of normalized counts across all samples
  • log2FoldChange — estimated effect size (log2 ratio group1/group2)
  • lfcSE — standard error of log2FoldChange
  • stat — Wald test statistic
  • pvalue — raw p-value
  • padj — Benjamini–Hochberg adjusted p-value (FDR)
• Optionally: normalized count matrices, VST/rlog-transformed matrices, PCA data, dispersion estimates, and .RData objects containing the DESeqDataSet and results objects.

---

8_wgcna/ — Weighted Gene Co-Expression Network Analysis

Weighted gene co-expression network analysis (WGCNA) output for the colony-level and patriline-level expression matrices. Contents typically include:

• Soft-threshold power selection tables and plots
• Adjacency and Topological Overlap Matrices (TOM)
• Module assignments (gene → module color)
• Module eigengene (ME) tables
• Module–trait correlation tables and p-values (traits include caste, tissue, colony, patriline, batch)
• Cluster dendrograms and module-detection outputs
• .RData session objects

Module colors (brown, turquoise, blue, yellow, green, grey, red) referenced in 17_cytoscape/ and 15_comparisons/wgcna/ are defined here.

---

9_rmats/ — Alternative Splicing (rMATS)

Top-level files

• GCA_043792635.1_ASM4379263v1_genomic.bed — BED coordinates of the reference genome features.
• splice_introns_GCA_043792635.1_ASM4379263v1_genomic.gtf — GTF annotation used for rMATS.
• introns_GCA_043792635.1_ASM4379263v1_genomic.agat.log — AGAT preprocessing log.
• as_counts.csv / as_counts_filt.csv — Alternative splicing event counts per gene, unfiltered and filtered.
• vespula_samples_and_traits.csv — sample-to-trait mapping used to define rMATS group files. See column descriptions in the 0_datasets/ section.
• 9_rmats_plots.RData — R workspace with objects used to generate splicing-related figures.

Subfolders

rmats_summaries/ — one <comparison>_summary.txt per pairwise comparison. Each summary reports counts of significant events per event type (SE/A3SS/A5SS/MXE/RI) at the standard rMATS thresholds.

rmats_genes_lists/ — one <comparison>.unique_genes_FDR0.05.txt per pairwise comparison. Plain text lists of the unique gene IDs harboring any significant splicing event at FDR ≤ 0.05.

rmats_outputs/<Factor>/<comparison>.tar.gz — full rMATS output archive per comparison, organized by factor (Batch/, Caste/, Tissue/, Patriline/, Colony/). Each expanded comparison directory contains the standard rMATS output set:

• summary.txt — event counts per type
• fromGTF.<EVENT>.txt — annotated events from GTF (all detected)
• fromGTF.novelJunction.<EVENT>.txt — novel junction events
• fromGTF.novelSpliceSite.<EVENT>.txt — novel splice site events
• JC.raw.input.<EVENT>.txt / JCEC.raw.input.<EVENT>.txt — raw junction (JC) and junction+exon (JCEC) count inputs
• <EVENT>.MATS.JC.txt / <EVENT>.MATS.JCEC.txt — main statistical results per event type

Column headers for the MATS.JC.txt / MATS.JCEC.txt tables include:

• ID — internal event ID
• GeneID, geneSymbol — gene identifiers
• chr, strand — genomic location
• Event-specific coordinate columns (e.g., exonStart_0base, exonEnd, upstreamES, upstreamEE, downstreamES, downstreamEE for SE; longExonStart_0base, longExonEnd, shortES, shortEE, flankingES, flankingEE for A5SS/A3SS; analogous coordinate sets for MXE and RI)
• IJC_SAMPLE_1, SJC_SAMPLE_1, IJC_SAMPLE_2, SJC_SAMPLE_2 — inclusion / skipping junction counts per replicate (comma-separated per sample)
• IncFormLen, SkipFormLen — effective lengths for inclusion / skipping isoforms
• PValue — likelihood-ratio p-value
• FDR — Benjamini–Hochberg adjusted p-value
• IncLevel1, IncLevel2 — per-replicate PSI values (Ψ, inclusion level) for the two groups
• IncLevelDifference — mean PSI difference (group1 − group2)

The tmp/ subdirectories contain intermediate per-event-type calculation files (rMATS_result_ID.txt, _INP.txt, _I-L.txt, _P-V.txt, _FDR.txt, _.txt) produced during the statistical model fit; most users will not need these.

---

10_tau/ — Tissue Specificity (Tau)

Tau index (Yanai et al.) computed from normalized expression across tissue conditions.

• normalized_counts.csv — normalized count matrix used for tau calculation. Rows = genes, columns = samples.
• tau.csv — per-gene tau values. Columns:
  • gene_id — gene identifier
  • tau — tissue specificity index (0 = broadly expressed; 1 = highly tissue-specific)
  • Additional columns may include per-tissue mean/median expression used in the calculation.
• tau_high.csv / tau_low.csv — subsets of genes classified as tissue-specific (high tau) vs broadly expressed (low tau) using the study's thresholds.
• tau_mean_tissue_specificity_tau.txt, tau_mean_tissue_specificity_tau_high.txt, tau_mean_tissue_specificity_tau_low.txt, tau_median_tissue_specificity_tau_high.txt, tau_median_tissue_specificity_tau_low.txt — summary statistics of tau across tissues (mean/median values within the high/low subsets).
• 10_tau.RData — R session object containing the tau computation environment.

---

11_emapper/ — eggNOG-Mapper Functional Annotation

Functional annotation of the reference proteome using eggNOG-mapper. All files are for GCA_043792635.1_ASM4379263v1.

• GCA_043792635.1_ASM4379263v1_protein.faa.gz — input protein FASTA.
• GCA_043792635.1_ASM4379263v1_protein.emapper.seed_orthologs.gz — diamond seed ortholog hits.
• GCA_043792635.1_ASM4379263v1_protein.emapper.hits.gz — raw HMM/diamond hits.
• GCA_043792635.1_ASM4379263v1_protein.emapper.annotations.gz — final functional annotation table. Columns follow the standard eggNOG-mapper v2 output:
  • query — input protein ID
  • seed_ortholog — best hit in eggNOG
  • evalue, score — alignment statistics
  • eggNOG_OGs — orthologous group assignments
  • max_annot_lvl — taxonomic level of best-annotated OG
  • COG_category — functional COG letter code(s)
  • Description — functional description
  • Preferred_name — recommended gene symbol
  • GOs — Gene Ontology terms (comma-separated)
  • EC — Enzyme Commission numbers
  • KEGG_ko, KEGG_Pathway, KEGG_Module, KEGG_Reaction, KEGG_rclass — KEGG identifiers
  • BRITE, KEGG_TC — additional KEGG functional hierarchies
  • CAZy — carbohydrate-active enzyme families
  • BiGG_Reaction — metabolic reactions
  • PFAMs — PFAM domain assignments
• GCA_043792635.1_ASM4379263v1_protein.emapper.decorated.gff.gz — GFF with annotations merged in as attributes.
• GCA_043792635.1_ASM4379263v1_genomic.map.gz — mapping between genomic feature IDs and protein IDs.

---

12_topgo_rrvgo/ — GO Enrichment and Semantic Reduction

topGO enrichment results and rrvgo semantic-similarity–reduced summaries for each gene list from earlier steps. This archive contains the primary per-list topGO tables; the 15_comparisons/go_terms/ directory contains the rrvgo-reduced versions for cross-comparison work (see below).

Subfolders: deseq2/colony/, deseq2/patriline/, rmats/, rrvgo/, tau/, topgo/, wgcna/, RData/

Typical topGO table columns:

• GO.ID — Gene Ontology term identifier
• Term — human-readable GO term name
• Annotated — number of genes in the background annotated to the term
• Significant — number of significant genes annotated to the term
• Expected — expected number of significant genes under the null
• classicFisher (or pvalue) — raw enrichment p-value (Fisher's exact test)
• weightFisher / elimFisher — p-values from topGO's weight/elim algorithms accounting for GO hierarchy
• Ontology (BP/MF/CC) is indicated by filename suffix.

---

15_comparisons/ — Integrated Cross-Analysis Comparisons

The largest and most complex archive. Integrates outputs across DESeq2, rMATS, WGCNA, and tau to identify overlapping and divergent genes/pathways across factors.

Top-level files

• binned_data_filtered.csv — filtered, binned expression matrix used as input to the integrated plots and hypergeometric tests.
• as_counts_filt.csv — filtered alternative-splicing event counts per comparison.
• Venn_numbers.csv — set sizes and intersection sizes used to draw Venn diagrams across gene lists (DEG / AS / WGCNA / tau).
• DEG_list.txt, AS_list.txt, WGCNA_list.txt — master gene lists per analysis type.
• 15_DEG_AS_plot.RData — R session object for the summary figure.

genes/ — Per-comparison significant gene lists (inputs)

Contains the gene lists that feed the integrated comparisons, mirroring the structure of the source analyses:

• genes/tau/tau_high.csv, genes/tau/tau_low.csv — tissue-specific vs broad genes.
• genes/deseq2/<design>/<Factor>/<Factor>.all.<groupA>_vs_<groupB>.csv — DEG lists, where <design> is colony or patriline and <Factor> is Batch, Caste, Tissue, Colony, or Patriline. Columns replicate the DESeq2 results structure (gene_id, baseMean, log2FoldChange, lfcSE, stat, pvalue, padj), typically filtered to padj ≤ threshold.
• genes/rmats/<Factor>/<groupA>.vs.<groupB>.unique_genes_FDR0.05.txt — plain-text lists of genes with any significant AS event (FDR ≤ 0.05).
• genes/wgcna/colony_wgcna_nodes.txt, genes/wgcna/patriline_wgcna_nodes.txt — gene-to-module assignments from WGCNA. Columns:
  • gene_id — gene identifier
  • module — WGCNA module color assignment
  • Additional columns may include module membership (kME), gene significance, and connectivity.

genes_output/ — Per-comparison overlap / hypergeometric matrices and RData objects

For each pairwise contrast, one or more matrices testing set overlaps among DEG, AS, and WGCNA-module gene lists:

• <Factor>.all.<groupA>_vs_<groupB>_deg_as_hyper_matrix.txt — hypergeometric-test matrix for DEG × AS overlap.
• <Factor>.all.<groupA>_vs_<groupB>_deg_wgcna_hyper_matrix.txt — DEG × WGCNA-module overlap.
• <Factor>.all.<groupA>_vs_<groupB>_deg_wgcna_as_hyper_matrix.txt — three-way DEG × WGCNA × AS overlap.
• <Factor>.all.<groupA>_vs_<groupB>_wgcna_as_hyper_matrix.txt — WGCNA × AS overlap.
• <Factor>.all.<groupA>_vs_<groupB>.RData — R session objects containing the underlying sets, contingency tables, and test results.

Typical columns in the hypergeometric matrices:

• set1, set2 — labels for the two gene sets compared (e.g., module name, comparison direction)
• set1_size, set2_size — number of genes in each set
• overlap — number of shared genes
• background_size — universe size used for the hypergeometric test
• pvalue — hypergeometric p-value
• padj — multiple-testing adjusted p-value
• odds_ratio or enrichment — fold enrichment of the observed overlap over expected

go_terms/ — rrvgo-reduced GO enrichment per gene list

Semantically clustered GO enrichment results (via rrvgo) for each comparison-derived gene list. Organized by source analysis and factor:

• go_terms/tau/rrvgo_tau_{high,low}_{BP,MF,CC}.csv
• go_terms/deseq2/<design>/<Factor>/rrvgo_<design>-<Factor>.all.<groupA>_vs_<groupB>_{BP,MF,CC}.csv
• go_terms/rmats/<Factor>/rrvgo_<groupA>.vs.<groupB>.unique_genes_FDR0.05_{BP,MF,CC}.csv
• go_terms/wgcna/<Design>/rrvgo_<design>_wgcna_nodes_{BP,MF,CC}.csv

Standard rrvgo output columns:

• go — GO term ID
• parent — ID of the representative "parent" term the row was clustered under
• parentTerm — human-readable name of the parent term
• term — human-readable name of this GO term
• score (often -log10(pvalue)) — enrichment strength used for clustering
• size — number of genes annotated to the term in the background
• cluster — cluster ID within the reduced set

go_terms_output/ — GO frequency summaries across comparisons

For each analysis type × factor × ontology, a frequency table counting how many pairwise comparisons a given GO term was significantly enriched in.

Files follow the pattern <source>_<Factor>_freq_<ONTOLOGY>.txt, e.g.:

• deseq2_colony_Colony_freq_BP.txt
• deseq2_patriline_Caste_freq_MF.txt
• rmats_Batch_freq_CC.txt
• wgcna_Colony_freq_BP.txt
• tau_high_freq_BP.txt, tau_low_freq_CC.txt

Columns:

• GO.ID — GO term identifier
• Term — human-readable term name
• Freq (or count) — number of comparisons in which the term was significantly enriched
• Ontology — BP, MF, or CC (also implied by filename)

Note: filename rmats_Patriline_freq_Mf.txt uses Mf (mixed case) rather than MF; treat as equivalent to the MF ontology.

---

16_correlations/ — Correlation Analyses

Correlation results (e.g., gene–gene, module–module, or module–trait Pearson/Spearman correlations) supporting the integrated interpretation. Typical columns in correlation tables:

• feature1, feature2 — labels for the two variables correlated (gene IDs, module colors, or trait names)
• correlation (or r, rho) — correlation coefficient
• pvalue — raw p-value
• padj — adjusted p-value
• n — number of observations

---

17_cytoscape/ — Cytoscape Network Exports

WGCNA-derived network files formatted for import into Cytoscape. Two sets of networks are provided:

• colony-* — networks derived from the colony-level WGCNA.
• patriline-* — networks derived from the patriline-level WGCNA.

Within each set, files are provided per module color (brown, turquoise, blue, yellow, green, grey, red) and at two connectivity levels:

• all — all nodes/edges in the module.
• top-30 — the 30 most-connected nodes/edges (hub genes) in the module.

File name pattern: <design>-CytoscapeInput-<all|top-30>-<nodes|edges>-<color>.txt

*-nodes-*.txt columns:

• nodeName — gene ID
• altName — alternate/annotated name (e.g., eggNOG-mapper preferred name)
• nodeAttr[nodesPresent] — module color(s) the gene belongs to

*-edges-*.txt columns:

• fromNode, toNode — connected gene IDs
• weight — topological overlap (TOM) or adjacency weight for the edge
• direction — edge direction (typically undirected)
• fromAltName, toAltName — annotated names for source/target nodes

---

Notes on Repeated Structures

Several patterns repeat across the archive and are described once here:

• DESeq2 result tables always share the same 6-column structure (baseMean, log2FoldChange, lfcSE, stat, pvalue, padj), regardless of which factor is contrasted.
• rMATS output directories contain the same file set for every pairwise comparison; only sample/group labels differ.
• rrvgo GO reduction tables share the same column schema across tau/, deseq2/, rmats/, and wgcna/ subfolders.
• Hypergeometric overlap matrices in 15_comparisons/genes_output/ follow one of a few templates (deg_as, deg_wgcna, wgcna_as, or three-way deg_wgcna_as); each has the same core columns described above.
• Direction pairs: for a given comparison, both A_vs_B and B_vs_A files typically exist. They contain the same underlying statistics; only the sign of log2FoldChange / IncLevelDifference and the ordering of columns is inverted.
• .RData files are R session snapshots providing programmatic access to the underlying analysis objects (recommended for reproducing figures or extending analyses).

---

Suggested Citation for Software Referenced

Users of this dataset should cite: DESeq2, WGCNA, rMATS-turbo, topGO, rrvgo, eggNOG-mapper v2, Cytoscape, and the tau tissue-specificity index as originally described by Yanai et al. (2005). Please refer to the associated manuscript for the primary citation and full methods.