Why this resource matters

Aquaculture breeding and management decisions increasingly rely on molecular evidence — which genes, proteins, or epigenetic marks track with an animal’s ability to survive heat events, disease outbreaks, acidification, or low salinity. That evidence already exists across dozens of published studies on Pacific oyster (Crassostrea gigas) and related shellfish. The problem is not a lack of data. It is that the data are scattered across different sequencing platforms, genome versions, phenotype definitions, and identifier systems, so no one can easily ask “which candidate genes show up consistently across independent thermal-tolerance studies?” without re-doing the comparison from scratch every time.

The proposal this resource implements describes the need directly:

“Develop standardized open-access, user-friendly, reproducible bioinformatics pipelines for resilience biomarker discovery through systematic reanalysis, data integration, and meta-analysis.”

The Aquaculture Resilience Evidence Engine (AREE) is a working system built to make that operational, not a literature summary or a static list of papers.

What this means for a breeder or researcher in practice

  • You can ask a specific question and get a traceable answer. Instead of reading six papers and mentally reconciling their different effect-size scales, sample sizes, and phenotype definitions, you can query AREE’s harmonized evidence table for a phenotype (say, thermal tolerance) and see every study’s contribution on a common scale, with the original study context preserved alongside it.
  • You can see convergence, and you can see disagreement. When two independent studies point at the same gene in the same direction, that is worth attention. When they point in opposite directions, AREE surfaces that conflict explicitly (see interpreting_meta_analysis.md) instead of quietly averaging it into a number that looks reassuring but hides real biological or methodological disagreement.
  • You can tell resilience evidence apart from stress-response noise. A gene that merely responds to heat is not the same as a gene that predicts which animals survive heat. AREE enforces this distinction structurally (see resilience_vs_exposure.md) so a candidate list isn’t quietly inflated with genes that just react to being stressed.
  • You get a transparent, auditable priority list, not a black box. Every candidate’s score is broken into named components you can inspect — number of supporting studies, direction agreement, effect size, phenotype relevance, and more — and no candidate is promoted to a high-confidence tier without meeting explicit replication and consistency requirements (see interpreting_candidate_scores.md). This matters for breeding programs: a candidate marker that will inform selection decisions needs a paper trail a skeptical geneticist can follow back to source data, not a single p-value from one study.
  • The system tells you what it doesn’t know. Missing metadata, uncertain identifier mappings, low replication, and processed-only data availability are recorded and surfaced, not hidden (see governance_and_provenance.md). A candidate list you can trust is one that is honest about its own weak points.

Why this is worth building now

Reanalyzing public aquaculture omics data one paper at a time does not scale, and every ad hoc reanalysis effort has to re-solve the same problems: which identifier system to standardize on, how to compare effect sizes across assay types, how to avoid conflating “changed under stress” with “predicts resilience.” AREE solves those problems once, as reusable infrastructure, so that adding the next public C. gigas dataset — or eventually a second shellfish species — is a matter of filling in a registration template and running a handful of commands, not re-inventing a harmonization pipeline.