The NEW TREND: genome-wide association studies (GWAS) with cohorts >1 million (or many millions of) individuals

These large multi-center consortia have the facilities to carry out such enormous GWAS, and the ability to collect all the necessary DNA (or more likely, to collect information from GWAS already completed, plus questionnaire data). Superficially, one might conclude that β€” simply performing the study β€” is not worthy of the data being published in a high-visibility journal, because there seems to be little chance these results might [a] improve prediction of risk, or [b] identify pathways that will “cure” or “modify the intensity of” these (very) complex disorders via development of novel drugs, etc.

Almost a decade ago, authors of more than one review concluded that “one could examine DNA samples from all 7.6+ billion people on this planet in one’s GWAS cohort, and β€” still β€” one could not approach finding all the small-effect genes contributing to the “variance explained” (in fact, probably not even 50% of the variance explained). However, we cannot be completely negative. It remains possible (but perhaps incomprehensible in today’s world?) that somewhere/sometime down the road, in a universe we cannot yet fathom, such data from these extremely large GWAS will be useful. And this is what keeps PI’s writing new grant proposals and publishing such large GWAS. πŸ™‚


From: Anonymous Ohio physician
Sent: Wednesday, March 6, 2019 4:36 PM

Wow…..another three more huge GWAS papers to read and attempt to comprehend before speaking. Thanks (I think). This reminds me of a definition I have provided to less sophisticated audiences for “genetics”…..”Chances are, if your parents did not have children, you won’t either”….. Seriously, the simple fact is that within our genomes lies some really great truths. Nevertheless, the knowledge breakthrough is important β€” if for no other reason than to shine a light on a few dark places.

From: Nebert, Daniel (nebertdw)
Sent: Wednesday, March 6, 2019 4:26 PM

On 25 Feb 2019, these GEITP pages described “an unprecedented cohort size of more than 1 million”, in which a genome-wide association study (GWAS) was carried out to search for genes (genotype; or genetic architecture) statistically significantly (P <5.0 x 10–8) associated with a phenotype (trait). In that Feb 25th publication, the complex trait being studied was insomnia; and these GEITP pages had predicted that β€” given the exploding numbers of humans whose DNA and personal history are being collected in repositories (not 'suppositories') β€” we will see this "trend of more-than-1-million individuals" in GWAS. Well, it has happened far sooner than expected. J Attached are THREE reports (all from the Feb 2019 issue of Nature Genetics) of GWAS β€” each one of which comprises a cohort exceeding 1 million individuals. The first [attached] study (in 'sample sizes up to 1.2 million subjects'), authors discovered 566 genetic variants in 406 genetic loci associated with multiple stages of tobacco use (initiation, cessation, and heaviness), as well as alcohol use, with 150 loci exhibiting evidence of pleiotropic association (i.e. a single gene that contributes to more than one trait). The second [attached] study (in a 'combined sample of over 1 million individuals'), authors studied several complex phenotypes ("general risk tolerance, adventurousness, and risky behaviors in the driving, drinking, smoking, and sexual domains") [!!]. Across all GWAS, authors identified hundreds of associated genetic loci β€” including 99 loci associated with general risk tolerance, shared genetic influences across risk tolerance and the risky behaviors β€” with bioinformatics analyses implying a role for glutamatergic and GABA-ergic neurotransmission (intriguingly, authors found no evidence of enrichment for genes previously hypothesized to be related to risk tolerance!!). In the third [attached] study, authors analyzed a large health insurance dataset to assess genetic and environmental contributions of 560 disease-related phenotypes in 56,396 twin-pairs and 724,513 sibling-pairs β€” out of 44,859,462 individuals living in the US. Authors estimated the contribution of environmental risk factors (socio-economic status, air pollution and climate) to each disease-related trait. They found significant heritability and shared environment for a substantial number of co-morbidities (i.e. the simultaneous presence of two or more chronic diseases or conditions in the same patient) and average monthly healthcare cost. Comprehending all the information in these GWAS, these days, is about the same as trying to drink water from a fire hose. πŸ™‚ DwN Nat Genet Feb 2o19; 51: 237-244 & 245-257 & 327-334

This entry was posted in Center for Environmental Genetics. Bookmark the permalink.