Common molecular methods regress phenotype on genome‑wide relatedness, assuming overall genetic similarity tracks the specific loci that cause a trait. Within‑family contrasts, assortative mating, non‑additive effects, and rare/structural variants can break this link, biasing estimates downward. The heritability 'gap' may be model misspecification, not missing biology.
— It warns policymakers and media not to treat low molecular heritability as proof against genetic influence when methods misalign with trait‑causal architecture.
Davide Piffer
2025.09.03
78% relevant
The article contrasts family/twin h² (~30–50% for fertility) with lower h²_SNP (~10–13%) and tiny h²_GWAS (~1–4%), and invokes 'hidden heritability' from cross-cohort heterogeneity (Tropf et al. 2017), aligning with the claim that common molecular methods and pooled models can underestimate true genetic influence.
Sebastian Jensen
2025.08.19
100% relevant
Jensen’s critique of GREML/RDR/sib‑regression assumptions and his sibling phenotype example despite higher autosomal sharing.
2025.05.14
65% relevant
The paper estimates heritability of four SES indicators using both family‑based and unrelated genotype‑based methods drawn from the same 170k‑person Norwegian cohort, offering a clean, within‑sample comparison of methods and showing estimates vary by method and trait—directly speaking to claims that molecular approaches can understate heritability relative to family designs.
2019.05.14
80% relevant
This paper uses a classical twin design and finds high heritability for lifetime earnings, aligning with the claim that family-based methods typically yield higher heritability than molecular approaches and warning against underestimating genetic influence from genome-average methods alone.
2018.01.08
60% relevant
The article notes that polygenic scores account for about 20% of intelligence’s ≈50% heritability, highlighting a persistent 'missing heritability' gap and the importance of method and sample-size advances—overlapping with the claim that molecular methods understate trait heritability.