“Heritability” is defined by “proportion of phenotypic variance (i.e. differences in any trait seen in a population) explained by genetic factors.” However, such estimates are uninformative with respect to the underlying genetic architecture [i.e. the underlying genetic basis of a phenotypic trait and ALL its variational properties. Phenotypic variation for quantitative traits (e.g. height, body mass index, BMI) is, at the most basic level, the result of segregation of alleles at quantitative trait loci]. Early on, with genome-wide association studies (GWAS; from 2006-16), it was first realized that heritability attributable to some common variants could be substantial. Then, as GWAS cohort sizes continued to increase — thereby identifying and including additional small-effect variants contributing increasingly smaller effects to the trait — the “revealed heritability” (or, ‘variance explained’) continued to grow, from 5-10% … but only to >20–25% for various traits and complex diseases.
At this point, many were suggesting that “including all humans on the planet as a cohort — still might not reach 100% heritability revealed.” GWAS in increasingly large human cohorts (recently these GEITP pages have noted several cohorts of >1 million individuals) have discovered thousands of single-nucleotide variants (SNVs) associated with various traits and complex diseases. To date, GWAS have mainly relied on arrays of common SNVs that are in linkage disequilibrium [LD*; the nonrandom association of alleles at different loci in a given population; these alleles need not always be ‘linked’ along the same chromosome!!! Remember that each gene has two alleles, one from each parent] with underlying causal variants.
Why is there still a gap between SNV data and pedigree heritability estimates? The most plausible hypotheses for this discrepancy are that: [a] causal variants are not well-tagged (or imputed) by common SNVs, because they are rare; and/or [b] that pedigree heritability
is over-estimated due to confounding effects with common environmental signals, or non-additive genetic variation. For “precision medicine” to improve, and to understand the associations between specific traits and fitness, considering the source of “still-missing heritability”, and achieving a better quantification of the genetic architecture of complex traits is important for experimental designs to map additional trait loci.
Authors [see attached article] address the hypothesis that the “still-missing heritability” is due to rare variants — not sufficiently tagged by common SNVs; they did this by estimating additive genetic variance for height and BMI from whole-genome sequence (WGS) data on a sample of 21,620 unrelated individuals of European ancestry. Authors assigned 47.1 million genetic variants to groups, based upon their minor allele frequencies (MAFs) and LD with variants nearby, and they estimated and partitioned the variations, accordingly. Estimated heritability increased dramatically to 0.79 for height and 0.40 for BMI, consistent with pedigree estimates.
Low-MAF variants in low LD with neighboring variants were enriched for heritability — and to a greater extent for protein-altering variants, consistent with negative selection thereon. Cumulatively, variants in the MAF range of 0.0001 to 0.1 explained 0.54 and 0.51 of heritability for height and BMI, respectively. These exciting findings [from the Jian Yang group (Institute for Advanced Research, Wenzhou Medical University, Wenzhou, Zhejiang, China) and the Peter M. Visscher group (Queensland Brain Institute, The University of Queensland, Brisbane, Queensland, Australia] therefore imply that the “still-missing heritability” of various traits and complex diseases — can be accounted for by rare variants, especially those in “regions of low LD“. 🙂
DwN
*I stumbled across a fantastic 2007 blog by Razib Khan (currently at UC Davis, CA?; it’s not clear that he has even obtained his PhD degree yet!!) on the complex topic of linkage disequilbrium. http://blogs.discovermagazine.com/gnxp/2007/01/basic-concepts-linkage-disequilibrium/#.XNRt0Y5Kg2w
This article is from the “preprint web site” (biorxiv.org), so it had not yet been accepted for publication, but I presume it will soon appear in Nature Genetics https://www.biorxiv.org/content/10.1101/588020v1