Multiple methods have been developed to estimate narrow-sense heritability, h 2 , using single nucleotide polymorphisms (SNPs) in unrelated individuals. However, a comprehensive evaluation of these methods has not yet been performed, leading to confusion and discrepancy in the literature. We present the most thorough and realistic comparison of these methods to date. We used thousands of real whole-genome sequences to simulate phenotypes under varying genetic architectures and confounding variables, and we used array, imputed, or whole genome sequence SNPs to obtain 'SNP-heritability' estimates. We show that SNP-heritability can be highly sensitive to assumptions about the frequencies, effect sizes, and levels of linkage disequilibrium of underlying causal variants, but that methods that bin SNPs according to minor allele frequency and linkage disequilibrium are less sensitive to these assumptions across a wide range of genetic architectures and possible confounding factors. These findings provide guidance for best practices and proper interpretation of published estimates.
Bibliographical noteFunding Information:
We thank D. Speed (University College London) for providing LDAK5. We thank the Keller and Vrieze lab groups, the Institute for Behavioral Genetics, N. Wray, A. Price, and S. Caron for helpful comments. This work was supported by NIH grant R01MH100141 (to M.C.K.), NHMRC grants 1078037 (to P.M.V.) and 1113400 (to P.M.V. and J.Y.), Sylvia & Charles Viertel Charitable Foundation Senior Medical Research Fellowship (to J.Y.), and NIH grants R01DA037904 and R01HG008983 (to S.V.). This work used the Janus supercomputer, which is supported by the National Science Foundation (award number CNS-0821794), the University of Colorado Boulder, the University of Colorado Denver,