Accurate Genomic Prediction of Human Height

Abstract

We construct genomic predictors for heritable but extremely complex human quantitative traits (height, heel bone density, and educational attainment) using modern methods in high dimensional statistics (i.e., machine learning). The constructed predictors explain, respectively, $40, 20, and 9% of total variance for the three traits, in data not used for training. For example, predicted heights correlate $0.65 with actual height; actual heights of most individuals in validation samples are within a few centimeters of the prediction. The proportion of variance explained for height is comparable to the estimated common SNP heritability from genome-wide complex trait analysis (GCTA), and seems to be close to its asymptotic value (i.e., as sample size goes to infinity), suggesting that we have captured most of the heritability for SNPs. Thus, our results close the gap between prediction R-squared and common SNP heritability. The $20k activated SNPs in our height predictor reveal the genetic architecture of human height, at least for common variants. Our primary dataset is the UK Biobank cohort, comprised of almost 500k individual genotypes with multiple phenotypes. We also use other datasets and SNPs found in earlier genome-wide association studies (GWAS) for out-of-sample validation of our results.

Influence of Genetic Interactions on Polygenic Prediction

“…Previous studies have mostly focused on improving parameter estimation, through increasing sample size and methodological improvement. For example, increasing sample size substantially increased accuracy of polygenic prediction of height within individuals of European ancestry (Lello et al 2018). Inclusion of samples of different backgrounds in the training data also helped (Martin et al 2019) ( Figure S2).…”

A Fast and Scalable Framework for Large-scale and Ultrahigh-dimensional Sparse Regression with Application to the UK Biobank

“…Recently Lello et al (2018) apply a lasso based method to predict height and other phenotypes on the UK Biobank. Instead of fitting on all QC-satisfied SNPs (as stated in Section 4.1), they pre-screen 50K or 100K most significant SNPs in terms of p-value and apply lasso on that set only.…”

“…While GWAS focus on identifying SNPs that may be marginally associated with the outcome using univariate tests, we would like to find relevant SNPs in a multivariate prediction model using the lasso. A recent study (Lello et al, 2018) fits the lasso to a similar subset of the dataset after one-shot univariate p-value screening and suggests improvement in explaining the variation in the phenotypes. However, the left-out variants with relatively weak marginal association may still provide additional predictive power in a multivariate environment.…”

“…The large models confirm the extreme polygenicity of standing height. Table 4: Comparison of prediction results on height with the model trained following the same procedure as ours except for an additional prescreening step as done in Lello et al (2018). In addition to R 2 , correlation between the fitted values and actual values is computed.…”

Developments in molecular epidemiology of aging

“…There is a range of machine learning analyses in use in molecular epidemiology. For example, lasso regression analysis of GWAS data is being used to develop prediction algorithms [69], although these have not yet been applied to studies of aging-related phenotypes; elastic-net regression analysis of epigenomic data is being used to develop novel measurements of the aging rate [24]; and neural-net analysis of a range of data types, including transcriptomic data, is being used to develop aging biomarkers and identify drug targets [70]. As machine learning approaches continue to mature, studies will be needed to compare methods and define best practices for implementation.…”