deextinction-feasibility

De-extinction feasibility

Treat “which dinosaur is most revivable, reconstruct its genome, and hatch it” as a sequence of testable evidence questions. Do not present the requested outcome as already feasible. Before comparing candidates, record exactly one objective: actual-biological-revival or comparative-proxy-research. Never put those objectives on one feasibility scale.

When comparing two or more candidates, call assess_deextinction_feasibility only with evidence classifications bound by the host to sealed ResearchRuns. The tool accepts no caller-authored score. Its soft research-evidence completeness is not a biological-feasibility score. PBDB coverage is shown separately and is excluded from completeness, gate decisions, and audit ordering; its biological-feasibility contribution is fixed at zero.

Required stages

1. Define the target. Distinguish biological resurrection of a non-avian dinosaur from engineering a phenotype proxy in an extant avian lineage. Record which outcome the researcher means.
2. Compare fossil candidates. Run analyze_paleontology_stratigraphic_support for each exact PBDB catalog, then pass 2–20 exact catalog-run plus analysis-run pairs to compare_fossil_candidate_evidence. The tool derives taxonomic resolution, returned versus available occurrences, truncation, formation/interval/georeference/primary-reference coverage, separate oldest/youngest bounds, interval width, and missingness. It accepts no score and produces no scalar, ordinal, Pareto, revival, or biological-feasibility ranking. Treat candidateOrdinal only as request order. Preserve the descriptive matrix when a parent is truncated, disclose truncated-parent-evidence, and retrieve missing pages before making count comparisons outside the tool.
3. Review molecular preservation. Search primary literature for specimen-specific, independently authenticated endogenous molecules. Record material, specimen identifier, geological age, authentication controls, contamination tests, replication, molecule type, and usable sequence length. Fossil abundance is not molecular preservation evidence.
4. Enrich museum holdings. Query current GBIF or museum records for specimen and collection candidates. Preserve institution, catalog number, basis of record, identifiers, license, and access date. A museum occurrence is not a sequence or morphology measurement.
5. Compare extant relatives. Use separately sourced and versioned extant avian/crocodilian genomes, explicit orthology rules, species tree, alignments, missingness, and uncertainty. Do not label an extant-relative consensus as a dinosaur genome.
6. Run hypothetical ASR. Only after the comparative-genomics inputs exist, estimate alternative ancestral states with posterior support and sensitivity to topology/model choice. Label every result hypothetical-ancestral-reconstruction; never recovered dinosaur sequence.
7. Assess development. Separate regulatory architecture, chromosome organization, cell-line engineering, germ-line transmission, host compatibility, embryogenesis, and viable hatching into distinct tests. Comparative sequence similarity alone supports none of them.
8. Handoff. Send the evidence ledger, negative findings, artifact receipts, unresolved questions, and stop-condition status to the built-in Research Director before advancing a stage or drafting a conclusion.