Review Papers Claim by Claim

 Quick Explanation Copied

 Long Explanation

What the paper does and how well it does it

Wojcik et al. leverage domestication as a replicated evolutionary experiment: 13 diploid crop species spanning 8 botanical families, 357 traits, standardized common-garden phenotyping, and 17–20 wild populations plus 16–22 landraces per species sampled near domestication centers to minimize genetic structuring and modern-breeding confounds Sampling and phenotyping designPreprint. The methodological core is elegant: Near-Infrared Spectra measured on leaves (NIRSleaf) serve as a control trait for non-domestication divergence (Pillai ≤0.16 vs ≥0.71 for true traits, apple excepted), enabling a corrected multivariate Phenotypic Divergence Index (mPDI) that is not inflated by trait count (Spearman p=0.57) NIRSleaf validation as control traitPreprint.

Key findings (reported observations): (1) Formal convergence of domestication traits shared by ≥3 species (resampling p=0.014), robust to Poaceae over-representation, with leaf length (8 species), leaf width (7) and seed area (9) most convergent — consistent with crop gigantism Crop gigantism reviewMilla & Matesanz 2017; (2) high phenotypic disjunction (mPDI mean 0.81) independent of domestication timing or mating system — supporting divergent selection over drift as the dominant force, consistent with predictions that domesticated forms are not a subset of wild space Grass domestication comparative reviewGlémin & Bataillon 2009; (3) phenotypic space shrinkage in 10/13 species; (4) progressive loss of variance–covariance matrix similarity with domestication age (Spearman p=0.004), implying reorganization of the G-matrix toward the anthropogenic adaptive landscape.

Critical assessment

Strengths: The convergence test is a genuine methodological advance — a design-matched resampling that respects the uneven trait matrix across species. Pre-registration-level transparency is excellent: data, R scripts, and MIAPPE trait ontology are openly deposited Open data and code availabilityPreprint. Limitations the authors partially address: NIRS PCoA axes are explicitly non-homologous across species (so SNIRS traits cannot be compared cross-species); maize mPDI is confounded by suboptimal greenhouse conditions; apple's near-zero mPDI stems from geographic sampling (Central Asia wild vs. European landraces), correctly diagnosed as drift during apple expansion Apple domestication genetics reviewCornille et al. 2014. Gaps: n=13 species with n≈40 accessions per form limits statistical power for mating-system tests (selfer effect p=0.15 is a tendency, not evidence); the mPDI conflates drift, selection, and evolvability — the authors acknowledge this is 'hard to decipher'; phenotypic correlations are used as proxies for G-matrices, which holds only under Cheverud's conjecture Phenotypic-genetic correlation proxyCheverud 1988; archaeological domestication timing estimates carry wide uncertainty that propagates into the timing–similarity regression. The negative space-size result (10/13 shrinkage) is inferential, not a formal significance test against a null of equal spaces.

What would falsify or extend the conclusions

Adding ~20 non-Poaceae species would test whether convergence survives family balance; measuring actual G-matrices (via pedigrees or genomic relatedness) in 2–3 species would verify the correlation-decoupling interpretation; direct tests of mPDI against fitness in human-altered habitats would validate the adaptive interpretation the authors propose Universal domestication definition frameworkLord et al. 2025. If mPDI were recomputed with wild populations sampled at matched distances from domestication centers and returned near zero, the disjunction conclusion would collapse.

Author-specific reviews

 BGPT Paper Review

Study Novelty

First formal cross-species statistical test of phenotypic convergence and the first multivariate divergence index (mPDI) with an internal control trait (NIRSleaf) — no prior study quantified domestication-space evolution across 8 botanical families with a shared framework.

Scientific Quality

Rigorous sampling, FDR control, design-matched resampling, open data/code, and explicit negative results (no timing or mating-system effects). Limitations: n=13 species, phenotypic-as-genetic proxy, non-homologous NIRS axes, maize greenhouse confound, apple sampling artifact — most acknowledged by authors.

Study Generality

Framework (mPDI, NIRS control) is explicitly portable beyond crops to local adaptation and experimental evolution; restricted mainly to plants since NIRS on homologous organs is impractical in animals.

Study Usefulness

Provides an immediately reusable comparative tool for domestication, local adaptation, and experimental evolution; the mPDI can rank species for conservation and pre-breeding prioritization of wild relatives.

Study Reproducibility

Data, spectra, R scripts, and MIAPPE trait ontology publicly deposited (doi 10.57745/QWEKVK; INRAE Forge); methods fully specified including spectral preprocessing parameters and random-skewers iteration counts.

Explanatory Depth

Connects observed patterns to theory (Lande's equation, random skewers, gene-network simulations, recombination dynamics); stops short of direct mechanistic tests — G-matrices inferred from phenotypic proxies only.