[study] LI2022_LOWSAL_VIBRIO (PRJNA756710)

[sr320]

Claimed by aree start-study on Hete, 2026-10-11. Run and harmonization finished 2026-10-11. Blocked on sr320/AREE#24, which supplies fetch-samplesheet --design-table. This branch contains that commit, so check-study-scope fails until #24 merges and this branch is rebased onto main.

Study

• study_id: LI2022_LOWSAL_VIBRIO. AREE's first real salinity evidence.
• Accession(s): BioProject PRJNA756710 (14 runs; 12 used, the two 0 h baselines not). The paper is Li et al. 2022, Front Immunol 13:859975, 10.3389/fimmu.2022.859975, open access; its data statement names this BioProject.
• Design: adult gill. Every oyster was injected with V. alginolyticus, then held at 10 ppt or 30 ppt and sampled at 12 h and 48 h. Three tanks per salinity; each library pools 3 oysters from one tank, so each library is a tank.
• Mode: raw_reanalysis, paired-end
• Machine and storage used: Hete (16 GB, 8 cores), native -profile local,workstation, Nextflow 26.04.7, data on /Volumes/Monarch (HFS+).
• Not in candidate_studies.md: chosen after its RNA-seq backlog ran out. Backlog [study] CIBNOR2021_HEAT_RRSS (PRJNA690951) #9 (PRJNA877226) is the same lab's high-temperature arm, with no control of its own, and should be marked rejected.

Result

Comparison	n (tanks)	padj < 0.05
10 vs 30 ppt, 48 h (meta_analysis_primary)	3 v 3	3,711 / 25,689
10 vs 30 ppt, 12 h	3 v 3	4,012 / 25,508

• Replicates are real: median within-group dispersion is 0.012–0.073.
• Agrees with the authors: 98.7% (12 h) and 99.0% (48 h) of their Supplementary Table 4 DEGs change in the same direction here; fold-change r = 0.87 and 0.85.
• The timepoints barely agree (r = 0.03 across all genes). The authors' lists show the same thing (74% same direction among genes DE at both), so it reflects biology, not the pipeline. It's recorded as the QC warning.
• Harmonized: 51,197 evidence records, 100% exact identifier mapping. Mapping 75.5–83.7%, unstranded (IU).

Checklist

• aree validate-study passes, with no warnings
• Sample sheet and FASTQ manifest generated by aree fetch-samplesheet. Salinity and timepoint appear only in the BioSample titles; design.tsv parses them, keeping the verbatim titles, and is joined on the deposited ID attribute. A test re-derives every row from its title and cross-checks it against the sample names.
• All 24 FASTQ MD5s verified against ENA before the run
• Workflow manifests committed; software_versions now reflect the native tools (from Build manifest software_versions from what actually ran #19)
• Standardized result tables and HTML reports committed; no raw reads, work/ or results/
• qc_status: passed_with_warnings, analysis_status: harmonized
• Limitations and flags recorded: both arms infected, sham not sequenced; mortality measured per group (ambiguous_phenotype_definition); pooled libraries; shared tanks across timepoints; design only in titles
• Curation tests in tests/test_li2022_lowsal_study.py (10; full suite 300 passed)
• No edits to shared code. Those are in Support single-end and 3' Tag-seq RNA-seq, and designs from authors' tables #24; rebase after it merges.

Curation decisions

• disease_susceptibility / disease (the ontology default), direction increase. Mortality over 12 d was 100% at 10 ppt vs 33% at 30 ppt among infected oysters, but measured per group; the sequenced animals' fate is unknown.
• Stressor salinity: both arms were infected, so salinity is the only factor that differs. The infection is recorded in stressor_original.
• 48 h prespecified as primary, before any result: it is past the acute osmotic shock, and the timepoints share tanks, so only one may pool.
• [study] CIBNOR2021_HEAT_RRSS (PRJNA690951) #9's possible control is deliberately not used. PRJNA877226 is plausibly this experiment's 33 °C arm (sample names C33-n vs C23-n here). Using this study's 30 ppt libraries as its control would cross BioProjects and possibly batches.

🤖 Generated with Claude Code