Pierre Fontanillas

Most lung cancers are tobacco related, with genetic factors influencing smoking behavior identified through genome-wide association studies. However, inherited risk in familial and non-smoking-related lung cancers, including risk in carriers of EGFR T790M, remains poorly understood. Here, in more than 3.3 million individuals, the EGFR T790M germline variant is significantly associated with lung cancer risk, with no increased risk for 17 other cancers and no interaction with polygenic risk. This risk exceeds that conferred by smoking and is several-fold higher in never-smokers. Global geographic and ancestry analyses show higher T790M prevalence in the US than in British- and Irish-descendant populations, reflecting a Southern Appalachian founder event about 200 to 225 years ago, increasing regional prevalence and affecting those of British, Irish, and African descent. Recognition of high-risk carriers may inform targeted genetic testing and screening strategies.

Background: Sex differences in alcohol use behaviors are well-established: males typically engage in heavier and more frequent drinking and exhibit more externalizing behaviors (e.g., other substance use), while females often transition to dependence more rapidly and present more internalizing psychopathology (e.g., depression). The biological mechanisms underpinning these differences are relatively unknown. Methods: In this study, we investigated the sex-differentiated genetic architecture of 11 alcohol use phenotypes pertaining to frequency, quantity and problematic use by leveraging sex-stratified genome-wide association studies (Ns 40,335 to 613,148). Specifically, we compared SNP-based heritability ( h ² _SNP ) estimates, individual genetic locus effects, genetic correlations ( r _g ) across alcohol phenotypes and with comorbid traits from independent GWAS, and polygenic score ( PGS ) associations with medical outcomes from clinical populations. Results: h ² _SNP was broadly similar between sexes, except for higher estimates in males for beer quantity and problematic alcohol use ( PAU ). We identified four sex-differentiated top loci ( p _sex-diff < 5 x 10 ^-8 ), including a female-specific association in IZUMO1 for drinking frequency and quantity , and three male-specific associations in ADH1B , KLB and FTO for beer quantity and/or PAU . Between-sex genetic correlations ranged from 0.68±0.07 to 0.89±0.04, these estimates were lowest for quantity measures and varied by beverage type, indicating partially distinct polygenic architecture. In males, we identified stronger positive genetic correlations with several externalizing traits (e.g., general addiction) compared to females. In females, we identified a specific positive genetic correlation with a single internalizing trait, self-harm. PGS analyses revealed sex-specific medical associations (e.g., bone/musculoskeletal conditions in females; hepatic/respiratory/infectious sequelae in males) that were obscured in sex-combined analyses; however, sex-specific PGS did not outperform combined-sex PGS for predicting alcohol use disorder diagnosis. Conclusions: Sex-aware analyses of alcohol use behaviors can improve our understanding of the genetic etiology of alcohol use and related health outcomes, and future studies should consider cultural variation (e.g., drinking attitudes, social norms) in the relationship between behavior and genetics.

Delay discounting (DD), a person’s preference for smaller immediate rewards over larger delayed rewards, is a heritable trait that is associated with psychiatric and physical outcomes, yet the biological mechanisms underlying these links are not known. We performed a GWAS of DD using 134,935 23andMe research participants and identified 11 genome-wide significant loci. We did not replicate our previously reported association with rs6528024 (chrXq13.3, GPM6B ; P = 5.30 × 10 ⁻⁰² ). The SNP-heritability of DD was 9.85 ± 0.57%. We observed genetic correlations between DD and 73 behavioral, physical, and neuroimaging traits, many of which persisted even after accounting for educational attainment, intelligence, and executive function. Network analysis revealed that the associations between DD and certain traits were explained by both overlapping and trait-specific biological processes. In a hospital-based cohort ( N = 66,917), DD polygenic scores were associated with 212 medical conditions. These results demonstrate that DD has a pleiotropic and polygenic common variant architecture, and is genetically associated with numerous outcomes, making it a promising endophenotype for psychiatric and physical health.

Cannabis is one of the most widely used drugs globally. We performed genome-wide association studies (GWASs) of lifetime (N = 131,895) and frequency (N = 73,374) of cannabis use. For lifetime cannabis use, we identified two loci, one near CADM2 (rs35827242, p = 4.63E-12) and another near GRM3 (rs12673181, p = 6.90E-09). For frequency of cannabis use, we identified one locus near CADM2 (rs4856591, p = 8.10E-09; r² = 0.76 with rs35827242). Lifetime and frequency of cannabis use were heritable (12.88 vs. 6.63%) and genetically correlated with previous GWASs of lifetime use and cannabis use disorder (CUD), as well as other substance use and cognitive traits. Polygenic scores (PGSs) for lifetime and frequency of cannabis use predicted cannabis use phenotypes in All of Us participants. A phenome-wide association study using a PGS for lifetime cannabis use to interrogate a hospital cohort replicated prior associations with substance use and mood disorders, and uncovered novel associations with celiac and infectious diseases. This work demonstrates the utility of pre-addiction phenotypes in cannabis use genomic discovery.

Literacy is a significant predictor of important life outcomes, such as attained education and income (Ritchie and Bates in Psychol Sci 24(7):1301–1308, 2013. 10.1177/0956797612466268) yet difficulties in reading and spelling are common. Both genetic and environmental factors account for individual differences in reading and spelling abilities (Little et al. in Behav Genet 47:52–76, 2017. 10.1007/s10519-016-9810-6), but there is some evidence that genetic factors can be moderated by environmental factors, many of which relate to differences in socio-economic status (SES). Studies in the US indicate that the heritability of reading and spelling abilities is higher in higher SES environments (Hart et al. in J Child Psychol Psychiatry 54(10):1047–1055, 2013. 10.1111/jcpp.12083; Friend et al. in Psychol Sci 19(11), 2008. 10.1111/j.1467-9280.2008.02213.x). Because countries differ in terms of factors such as education access and social mobility, the genetics (or simply gene) x SES interaction may or may not be present in other populations. Here, we utilise summary statistics from a well-powered genome-wide association study on dyslexia (Doust et al. in Nat Genet 54:1621–1629, 2022. 10.1038/s41588-022-01192-y) to construct polygenic indices in two cohorts of children/adolescents in Australia (N = 1315) and the United Kingdom (N = 5461 at age 7; N = 4306 at age 16), and test whether the effect of measured genes on variation in reading ability is moderated by family SES. While polygenic indices and SES both showed statistically significant effects on reading and spelling performance, no interaction effect was found. These results are contrary to results of some twin studies in the United States that have found an interaction effect. Yet, these findings support the broader literature on gene x SES interaction that mostly report no such interaction in other cognitive traits outside the United States suggesting country differences in how strongly SES relates to education quality.

The ability to read is an important life skill and a major route to education. Dyslexia, characterized by difficulties with accurate/ fluent word reading, and poor spelling is influenced by genetic variation, with a twin study heritability estimate of 0.4–0.6. Until recently, genomic investigations were limited by modest sample size. We used a multivariate genome-wide association study (GWAS) method, MTAG, to leverage summary statistics from two independent GWAS efforts, boosting power for analyses of dyslexia; the GenLang meta-analysis of word reading (N = 27,180) and the 23andMe, Inc., study of dyslexia (N_cases = 51,800, N_controls = 1,087,070). We increased the effective sample size to 1,228,832 participants, representing the largest genetic study of reading-related phenotypes to date. Our analyses identified 80 independent genome-wide significant loci, including 36 regions which were not previously reported as significant. Of these 36 loci, 13 were novel regions with no prior association with dyslexia. We observed clear genetic correlations with cognitive and educational measures. Gene-set analyses revealed significant enrichment of dyslexia-associated genes in four neuronal biological process pathways, and findings were further supported by enrichment of neuronally expressed genes in the developing embryonic brain. Polygenic index analysis of our multivariate results predicted between 2.34–4.73% of variance in reading traits in an independent sample, the National Child Development Study cohort (N = 6410). Polygenic adaptation was examined using a large panel of ancient genomes spanning the last ~15 k years. We did not find evidence of selection, suggesting that dyslexia has not been subject to recent selection pressure in Europeans. By combining existing datasets to improve statistical power, these results provide novel insights into the biology of dyslexia.