Abstract
Background Lung adenocarcinoma (LUAD) among never-smokers is a public health burden especially prevalent in East Asian (EAS) women. Polygenic risk scores (PRSs), which quanefy geneec suscepebility, are promising for straefying risk, yet have mainly been developed in European (EUR) populaeons. We developed and validated single-and mule-ancestry PRSs for LUAD in EAS never-smokers, using the largest available genome-wide associaeon study (GWAS) dataset.
Methods We used GWAS summary staesecs from both EAS (8,002 cases; 20,782 controls) and EUR (2,058 cases; 5,575 controls) populaeons, as well as independent EAS individual level data. We evaluated several PRSs approaches: a single-ancestry PRS using 25 variants that reached genome-wide significance (PRS-25), a genome-wide Bayesian based approach (LDpred2), and a mule-ancestry approach that models geneec correlaeons across ancestries (CT-SLEB). PRS performance was evaluated based on the associaeon with LUAD and AUC values. We then esemated the lifeeme absolute risk of LUAD (age 30-80) and projected the AUC at different sample sizes using EAS-derived effect-size distribueon and heritability esemates.
Findings The CT-SLEB PRS showed a strong associaeon with LUAD risk (odds raeo=1.71, 95% confidence interval (CI): 1.61, 1.82) with an AUC of 0.640 (95% CI: 0.629, 0.653). Individuals in the 95th percenele of the PRS had an esemated 6.69% lifeeme absolute risk of LUAD. Comparison of LUAD risk between individuals in the highest and lowest 20% PRS quaneles revealed a 3.92-fold increase. Projeceon analyses indicated that achieving an AUC of 0.70, which approaches the maximized prediceon poteneal of the PRS given the esemated geneec variance, would require a future study encompassing 55,000 EAS LUAD cases with a 1:10 case-control raeo.
Interpretations Our study underscores the poteneal of mule-ancestry PRS approaches to enhance LUAD risk straeficaeon in never-smokers, parecularly in EAS populaeons, and highlights the necessary scale of future research to uncover the geneec underpinnings of LUAD.
Introduction
Lung cancer is a major global health challenge, responsible for about 20% of all cancer deaths in 20201. While smoking is the primary eeologic factor, around 25% of lung cancer cases occur in never-smokers, with significant geographical variaeons2 . Notably, in East Asia (EAS), never- smoking women exhibit high incidences of lung adenocarcinoma (LUAD), the predominant histologic subtype3.
Beyond smoking, risk factors for LUAD include environmental and occupaeonal exposures, lifestyle, family history, and geneec suscepebility4–7. Genome-wide associaeon studies (GWAS) for never-smoking lung cancer across EAS and European (EUR) populaeons have idenefied 28 suscepebility variants at 25 independent loci8–13. These GWAS findings enhance our understanding of LUAD’s geneec architecture, yet the translaeon of these findings into clinical pracece requires further invesegaeon.
Polygenic risk scores (PRS) aggregate the effects of individual single nucleoede polymorphisms (SNPs) to esemate geneec risk for diseases, playing a crucial role in cancer preveneon and screening. Previous efforts primarily uelized EUR populaeons for lung cancer PRS development14,15 to beVer idenefy high-risk individuals. Nonetheless such PRSs omen fall short for EAS never-smokers, refleceng a bias in screening guidelines and the focus of studies on populaeons with EUR ancestry. Despite ongoing efforts to integrate PRS into clinical praceces for complex diseases, as seen in inieaeves like the electronic Medical Records and Genomics (eMERGE) network16, Veterans Affairs Genomic Medicine at Veteran Affairs (GenoVA) study17, and the Women Informed to Screen Depending On Measures of risk (WISDOM) study18, lung cancer has been excluded. The predominant global burden of lung cancer, driven primarily by smoking, overlooks the significant impact and poteneal benefits of PRS in never-smoking populaeons.
The recent expansion of GWAS across diverse populaeons, combined with advancements in PRS methodology, underscores a pivotal shim towards enhancing disease outcome prediceon beyond EUR populaeons19. Notably, the development of mule-ancestry PRS methods marks a considerable advancement, leveraging data from various ancestral backgrounds to enrich the prediceve accuracy and robustness of PRSs20–26. By jointly modeling geneec data from muleple populaeons, these approaches improve the prediceve power of PRSs, thereby enhancing disease outcome prediceons in non-EUR populaeons.
While the development and validaeon of PRSs are criecal, the applicaeon of an established PRS in esemaeng the absolute risk of a disease offers valuable insights into risk straeficaeon, poteneally guiding clinical interveneons, such as lung cancer screening. Moreover, projeceng necessary sample sizes for future studies to achieve varying levels of PRS prediceve accuracy is an underexplored area that could significantly impact epidemiological research design and cost- effeceveness.
This study aims to evaluate the prediceve performance of single- and mule-ancestry PRSs for LUAD in never-smoking EAS individuals using state-of-the-art methodologies. By integraeng muleple datasets, we constructed PRSs based on summary staesecs for 5,622 never-smoking cases and 21,813 never-smoking controls from EAS and EUR ancestries. We assessed the performance of the PRSs using independent, individual level data of 4,438 never-smoking EAS cases and 4,544 never-smoking EAS controls. Furthermore, we esemated the lifeeme and 10- year absolute risks of LUAD using the most accurate PRSs developed. Lastly, we projected the sample sizes needed in future research to achieve specific levels of prediceon accuracy with PRSs in EAS never-smokers (Figure 1).
Summary staesecs from East Asian (EAS) genome-wide associaeon studies (GWAS) were used to develop single-ancestry PRS using methods such as a simple PRS constructed using 25 SNPs that have previously reached genome-wide significance (i.e., P<5x10-8) (PRS-25), a PRS using the clumping and thresholding (CT) method (PRS-CT) incorporaeng 8 SNPs, and a PRS using a genome-wide Bayesian-based approach, LDpred2 (LDpred2 PRS) incorporaeng close to a million SNPs. For the mule-ancestry PRS development, we also used summary staesecs from European (EUR) GWAS, applying the PRS-CSx method that leveraged genome-wide associaeon summary staesecs for close to a million SNPs with a Bayesian conenuous shrinkage prior to model SNP effect sizes across populaeons, as well as CT-SLEB method, which enhances the standard CT methods with a two-dimensional approach to select SNPS for EAS PRS construceon by incorporaeng over 2 million SNPs. Tuning and validaeon of each PRS was conducted in an independent EAS individual-level data. Relaeve risk per PRS quanele was calculated as an odds raeo (OR) with the middle quanele (40th to 60th percenele) set as the reference, and the area under the receiver operaeng curve (AUC) was esemated for each PRS. CT-SLEB PRS was used to esemate 10-year and lifeeme cumulaeve absolute risk, and PRS-CT and PRS-LDpred2 were used for sample size projeceon.
Methods
Study design and data sources
Training data compilaeon: study populaeon and genotyping
The studies, genotyping protocols, and quality control for PRS construceon are previously described in detail27. Briefly, EAS single-ancestry PRSs for LUAD were constructed using summary data from 3,564 never-smoking LUAD cases and 16,238 never-smoking controls of EAS ancestry from the Nanjing Lung Cancer Study (NJLCS)9,28, Naeonal Cancer Center of JAPAN (NCC), and the Research Insetute and Aichi Cancer Center (ACC). The NJLCS study combined data from several ciees, genotyped by Affymetrix Genome-Wide Human SNP Array 6.0, as well as the Oncoarray GWAS28–30. The NCC study included lung cancer paeents and control data from muleple Japanese, with genotyping conducted on Illumina HumanOmniExpress and HumanOmni1-Quad genotyping plaqorms. Similarly, the ACC study pooled data from the muleple Japanese medical insetueons, and the Nagahama Study, with genotyping performed using Illumina 610k and Illumina660k plaqorms. To construct the mule-ancestry PRS, we further incorporated GWAS summary staesecs from 2,058 never-smoking LUAD cases and 5,575 never- smoking controls of EUR ancestry, genotyped using Illumina Infinium OmniExpress-24 v1.2 BeadChips and Illumina Human660W-Quad BeadChip13.
Tuning and validaeon: study populaeon and genotyping
For PRS tuning and validaeon, we used 4,438 never-smoking LUAD cases and 4,544 never- smoking controls from the Female Lung Cancer Consoreum in Asia (FLCCA)10,11, an internaeonal consoreum composed of never-smoking EAS women from regions including Mainland China, Singapore, Taiwan, Hong Kong, South Korea and Japan 10,11. All samples were genotyped using the Illumina 660W, 370K or 610Q microarrays. The FLCCA data were randomly and equally divided for tuning (2,219 cases, 2,272 controls) and validaeon (2,219 cases, 2,272 controls).
Informed consent was obtained from all parecipants, and the recruitment and data colleceon procedures were approved by the local ethics review commiVees of all parecipaeng research insetutes.
Construction of PRS
Single-ancestry PRS methods
PRS-25: We constructed a simple PRS using the 25 independent SNPs (PRS-25) that reached genome-wide significance (i.e., P<5x10-8) in the largest GWAS for LUAD in EAS populaeon to date8. The PRS-25 was calculated by summing the risk alleles weighed by their effect sizes (i.e., per allele log-odds raeo) obtained from a meta-analysis including only never-smokers.
PRS-CT: We generated a PRS using the CT method31 using PLINK 1.9032. This involved the clumping of SNPs based on LD with a threshold of r2=0.1 within a 500kb window. The LD were esemated from half of the FLCCA dataset used for tuning dataset. We created nine SNP subsets by applying incremental P-value thresholds (i.e., P<5x10-8, P<5x10-7,…, P<5x10-1, 1), and calculated their respeceve PRSs using PLINK232 command “--score cols=+scoresums,-scoreavgs no-mean-imputaeon”. Using the tuning dataset, the AUC was calculated for each threshold to idenefy the most prediceve P-value cutoff.
LDpred2-PRS: We applied the LDpred2 method, implemented in the R package bigsnpr33, using a Bayesian framework to esemate SNP effect sizes through a shrinkage esemator. This method leverages GWAS summary staesecs, incorporaeng a prior for effect sizes, while also accouneng for LD across SNPs. Our analysis was constrained to HapMap3 variants, and we calibrated the model using a range of hyperparameters: the proporeon of causal SNPs was set across a 21- point logarithmic sequence from 10-5 to 1, and the per-SNP heritability was set as a fraceon (0.3, 0.7, 1 or 1.4) of the total heritability esemated by LD score regression34 divided by the number of causal SNPs. Lastly, we used the “sparse” opeon to set weaker effects to zero. The opemal tuning parameters were selected based on the highest AUC achieved on the tuning dataset.
PRS-EUR 128: We evaluated the performance of a European-derived PRS (PRS-EUR 128)35, which consists of 128 variants obtained from a GWAS of predominately aceve smokers, on EAS populaeons to assess its cross-populaeon applicability.
Mule-ancestry PRSs methods:
LDpred2 PRS + PRS-EUR 128: For the mule-ancestry PRSs, we applied the weighted-PRS approach by linearly combining the most prediceve single-ancestry PRS from the EAS populaeon (Ldpred2 PRS) with the EUR populaeon-specific PRS (PRS-EUR 128). The weights of this combined PRS were calculated by applying a logisec regression on the tuning dataset in R version 4.2.0.
PRS-CSx: We applied PRS-CSx, a mule-ancestry polygenic prediceon method that uses a Bayesian framework with a conenuous shrinkage prior to esemate SNP effect sizes from GWAS summary staesecs across different populaeons26. LD reference panels for EUR and EAS, provided by the PRS-CSx somware, were constructed from the 1000 Genomes Project (1KG) samples 36. We used the default sewng with the gamma-gamma prior hyperparameters (a and b) at 1 and 0.5, respecevely. The shrinkage parameter 𝜙 was assessed at 10-6, 10-4, 10-2, and 1 within the tuning dataset to select the value with highest AUC. With the opemal 𝜙, we calculated the PRS using posterior effect-sizes from both EAS and EUR populaeons. Weights were then esemated to linearly combine the two PRSs on the tuning dataset. The final performance was then evaluated on an EAS independent validaeon dataset.
CT-SLEB: We used the recently developed CT-SLEB method to derive a PRS using EAS and EUR datasets 24. This method extends the standard CT method by designing a two-dimensional approach to select SNPs for EAS PRS construceon. It incorporates an empirical Bayesian (EB) framework to model geneec correlaeons between EAS and EUR. Following this, a super-learning (SL) model is then applied to integrate muleple PRSs, each generated using disenct p-value thresholds and clumping parameters37. Our implementaeon of CT-SLEB followed the default sewng with p-value thresholds 𝑝! = 5x10"#, 5x10"$, …, 5x10"%, or 1, and geneec distances 𝑑 = 50/𝑟& or 100/𝑟&, where 𝑟& = 0.01, 0.05, 0.1, 0.2, 0.5 𝑜𝑟 0.8. The clumping process for EAS was based on the LD reference derived from the FLCCA tuning dataset, and the EUR LD reference was based on the 1KG EUR samples.
Relative and absolute risk calculation for PRSs
To evaluate each PRS’s performance, we standardized the scores to a unit standard deviaeon, based on the control group in the EAS validaeon dataset. We then quanefied the associaeon between the standardized PRSs and LUAD risk using the OR and 95% CI, via logisec regression, adjuseng for age and the first 10 principal components (PCs). We evaluated the prediceve performance of the different PRS models through the adjusted AUC values, accouneng for age and top 10 PCs, using the R package RISCA37. We further evaluated the ORs of each PRS for risk of LUAD based on PRS perceneles, sewng the middle quanele (40-60%) as the reference category.
For absolute risk esemates, we used the iCARE somware38 to calculate the cumulaeve lifeeme (age 30-80) and 10-year absolute risks of LUAD among never-smoking female controls in FLCCA (N=4,544). Absolute risks were derived by applying the Cox proporeonal hazard model with the top performing PRS (CT-SLEB) and age-specific lung cancer incidence and mortality rates in Taiwan39.
Projections of CT and LDpred2 PRS performance by sample size
We used GENESIS package 40 to esemate the sample sizes needed for PRS to reach various AUC levels in the EAS populaeon. This method esemates the expected number of SNP discoveries and their explained heritability in future studies. Using the GWAS summary staesecs from our EAS training dataset, and the provided LD scores for EAS populaeons from the 1KG dataset, we projected the AUC for CT PRS across various case-control raeos, from 1:1 to 1:10, and case numbers ranging from 5,000 to 200,000. Given GENESIS’s specific design for CT PRS, we developed a extend its applicaeon to LDpred2 PRS projeceons. This involved modeling the relaeonship between effeceve sample sizes with the phenotypic variance raeo between LDpred2 and CT PRS (Supplementary Figure 1), elaborated in the Supplementary Note.
Results
Development and validation of the PRS
We applied several cuwng-edge single-and mule-ancestry PRS methods (Methods, Supplementary Figure 2), evaluaeng their performance in terms of relaeve risk and AUC within the EAS validaeon dataset (Methods, Table 1, Supplementary Figure 2-3). Among single- ancestry PRS methods, the LDpred2 PRS, incorporaeng 942,591 SNPs, outperformed other methods, with an odds raeo (OR) per unit standard deviaeon (SD) of 1.62 (95% confidence interval (CI): 1.52, 1.73) alongside an adjusted AUC of 0.629 (95% CI: 0.618, 0.641). In contrast, the PRS designed exclusively for EUR populaeons underperformed in our EAS populaeon, with an esemated adjusted AUC of 0.489 (95% CI: 0.477, 0.501), possibly aVributed to the inclusion of smokers in the development of the exiseng EUR PRS.
Prediction performance of different methods for generating polygenic risk scores for lung cancer in never-smoking East Asian populations.
Among mule-ancestry PRS methods, the weighted-PRS of EAS and EUR yielded an OR per unit SD of 1.62 (95% CI: 1.52, 1.72), with PRS-CSx and CT-SLEB showing even stronger associaeon with ORs of 1.67 (95% CI: 1.57, 1.78) and 1.71 (95% CI: 1.61, 1.82), respecevely. Among all the PRS methods, CT-SLEB had the best performance in the validaeon dataset with an AUC of 0.640 (95% CI: 0.629, 0.653). Sewng individuals in the middle PRS quanele (40th to 60th percenele) as the reference category, those in the highest 5% of risk for the top performing PRS, CT-SLEB, had
4.17 (95% CI: 3.20, 5.47)-fold risk, whereas those in the lowest 5% had 0.33 (95% CI: 0.21, 0.51)- fold risk of developing LUAD (Figure 2).
Odds raeos of PRS per standard deviaeon (SD) and 95% confidence intervals are shown for the single-ancestry 25 SNP polygenic risk score, PRS-25 (A), Clumping and thresholding method, PRS-CT (B), Bayesian-based genome-wide approach, LDpred2 PRS (C), mule-ancestry approach, CT-SLEB (D).
Absolute risk of developing lung adenocarcinoma by CT-SLEB PRS
Using the iCARE package38, we esemated the absolute risk of LUAD for never-smoking EAS women, uelizing the CT-SLEB PRS which demonstrated the highest AUC (Table 1). The cumulaeve lifeeme risk of LUAD, assessed between age 30-80, varied markedly across PRS perceneles, ranging from a minimal 0.78% in the 5th percenele to a substaneal 6.69% in the 95th percenele (Figure 3). Addieonally, the 10-year absolute risk for LUAD in a never-smoking 50-year-old woman, a criecal age for inieaeng recommended annual lung cancer screening as per the United States Preveneve Services Taskforce (USPSTF), varied from 0.13% in the 5th percenele to 1.11% in the 95th percenele of the PRS distribueon, emphasizing the PRS’s poteneal in guiding lung cancer screening strategies.
Lifeeme Cumulaeve (A) 10-year (B) absolute risk of developing lung adenocarcinoma in never-smoking women in East Asia by perceneles of the CT-SLEB polygenic risk score (PRS). Absolute risks were calculated based on Taiwan’s age specific incidence and mortality data and using the PRS relaeve risks esemated as described in the material and methods seceon.
Projections of polygenic risk score performance by sample size
Using the GENESIS model, we projected the expected AUC of PRS-CT and LDpred2 PRS under varying GWAS sample sizes and case-control raeos for LUAD in never smoking EAS women (Figure 4). Our analysis, using the EAS training dataset, esemated approximately 1,772 (s.e.=1,641) suscepebility variants that are independently associated with LUAD27. This high number of variants underscores the extensive polygenic nature of LUAD, implying relaevely small effect sizes for individual SNPs. Based on our current data, the expected AUC for the LDpred2 PRS stands at 0.631 (95% CI: 0.618, 0.641), aligning closely with the actual performance of the LDpred2 PRS in the validaeon dataset. As our projeceon increases in sample size to 65,000 cases with a one-to-one case-control raeo and a one-to-ten case-control raeo, the AUC is predicted to rise to 0.681 (95% CI: 0.601, 0.726) and 0.704 (95% CI: 0.629, 0.728), respecevely. The upper boundary of our prediceons, based solely on PRS, suggests an AUC poteneal of up to 0.731 (95% CI: 0.640,0.786).
The expected area under the receiver operating characteristic curve (AUC) of a polygenic risk score (PRS) built based on a GWAS of specified sample sizes for lung adenocarcinoma in never-smoking East Asian women. Projected AUC values for PRS-CT and LDpred2 PRS with a 1:1 and a 10:1 case to control raeo (A) and AUC values for LDpred2 PRS for 1:1, 2:1, 4:1, 5:1 and 10:1 case to control raeo (B).
Explaining gene1c variance through PRS across different sample sizes
We also evaluated the geneec variance explained by the PRS under different sample sizes given the projected results (Supplementary Note, Table 2). The current LDpred2 PRS explained 26.6% of the genome-wide chip heritability, contribueng to approximately 16.5% of the 1.84-fold familial relaeve risk for lung cancer among EAS never-smokers. Concurrently, genome-wide chip variants heritability explained 61.9% of the 1.84-fold familial relaeve risk associated with the disease. With an expanded sample size of 35,000 cases and 350,000 controls, the constructed PRS is projected to account for 57.9% of the genome-wide chip variants heritability, and 35.8% of the 1.84-fold familial relaeve risk for the disease.
Genetic variance in East Asian lung cancer among never smokers explained by LDpred2 PRS
Discussion
We developed and validated single- and mule-ancestry PRSs for LUAD in never-smoking EAS individuals using the largest GWAS dataset of never-smokers to date. The mule-ancestry PRS method, CT-SLEB, integraeng summary data from EAS and EUR never-smokers, emerged as the best-performing PRS. It exhibited a dose-response relaeonship with LUAD risk and achieved higher AUC than all other evaluated PRSs. Further, our analysis demonstrates the poteneal of PRS in straefying individuals’ 10-year and lifeeme risk of developing LUAD. Lastly, we projected the expected discriminatory accuracy of the PRS across a range of sample sizes and case-control raeos.
To date, GWAS and subsequent PRS models have largely centered on EUR populaeons 41, 42, rendering them less precise when applied to non-EUR populaeons and risking the exacerbaeon of health dispariees 43, 44. Consistent with prior research, our study observed that a lung cancer PRS constructed using EUR data significantly underperformed when applied to EAS individuals15. However, by uelizing novel methods to integrate GWAS data from both EUR and EAS populaeons, we demonstrated that a mule-ancestry PRS enhances the precision of risk straeficaeon for LUAD among EAS never-smokers.
While large-scale efforts to integrate PRS into clinical pracece have focused on condieons with well-established PRS prediceon performance16–18, such as Type 2 Diabetes, breast cancer, and cardiovascular diseases, lung cancer has been notably absent from these efforts, primarily due to the overwhelming focus on smoking as a risk factor. This has limited the exploraeon and applicaeon of PRS in lung cancer risk straeficaeon, parecularly among never-smokers. Our study addresses this gap by preseneng a PRS that signifies a 3.92-fold increase in lung cancer risk for individuals in the top 20% risk quanele, a marked improvement over the 2.09-fold increase observed in the prior study by Wei et al. 45 for a cohort of Chinese never-smoking women.
The CT-SLEB method’s improved performance can be aVributed to its ability to uelize diverse genotyping arrays, thereby including populaeon-specific variants. This aspect, along with the inclusion of EUR samples in the training set, broadens the geneec diversity and leverages larger sample sizes for improved model accuracy. Moreover, CT-SLEB excels in modeling geneec correlaeons across ancestries, efficiently using both shared and unique geneec markers for refined disease risk esemaeons across populaeons. Lastly, we set up our study with a robust three-sample split design, dividing datasets into training, tuning, and validaeon. This ensures that PRS efficacy is validated independently, effecevely reducing the overfiwng risk.
Current PRS studies predominantly focus on evaluaeng relaeve and absolute risks8,12, 15, 45, 46, yet our research introduces a crucial forecast of sample sizes required for differing levels of PRS accuracy. Our projeceon analyses indicated that achieving an AUC of 0.70 would approach the maximized prediceon poteneal of the PRS, given the esemated geneec variance explained by GWAS chip variants. To reach this level of accuracy, a future study would need to include 65,000 EAS LUAD cases with a 1:4 case-control raeo, or 55,000 cases with a 1:10 case-control raeo.
Notably, accessing large biobanks with publicly available controls could reduce the number of required cases47. These projeceons offer a strategic framework for planning and designing future geneec studies on lung cancer, establishing clear benchmarks for PRS model performance, which is esseneal for the scienefic community in advancing geneec risk prediceon.
Our study has several limitaeons. First, it focuses on geneec suscepebility without considering addieonal queseonnaire-based risk factors, aiming primarily to idenefy the most effeceve PRS model within geneec contexts. We plan to conduct integraeve analyses that merge PRS with other risk indicators to refine LUAD risk prediceons for never-smokers. Second, our projeceons are tailored to single-ancestry PRS models, not fully addressing the intricacies of mule-ancestry approaches, including diverse sample sizes and geneec correlaeons across populaeons. Third, our validaeon cohort, derived from the FLCCA and covering various EAS regions, relied on Taiwanese incidence data for absolute risk esemaeons. Consequently, our absolute risk findings may not extend universally to all EAS never-smokers, though our PRS performance assessment and sample size projeceons remain applicable to our populaeon of interest. Lastly, the clinical applicaeon of our PRS findings, parecularly in screening and risk counseling, is sell in preliminary stages. Despite advancements in using PRS for smoking cessaeon trials48, its implementaeon for never-smokers in clinical sewngs has been limited, preseneng a significant area for future research.
In summary, our study evaluates various PRS models to capture the geneec predisposieon to LUAD among never-smoking EAS individuals. It extends beyond risk prediceon by esemaeng both 10-year and cumulaeve lifeeme absolute risks, and by projeceng the sample sizes required for future GWAS to refine the prediceve power of PRSs in future GWAS. Addieonally, we quanefy the phenotypic variance captured by PRSs across different sample sizes. Future studies are crucial to further improve these PRS models, aiming to enhance geneec risk prediceons while integraeng a wider array of risk factors. Such efforts will develop more accurate and comprehensive risk models for LUAD in never-smoking individuals across diverse populaeons.
Data Availability
All data supporting the findings described in this paper are available in the paper and in the Supplementary Information and from the corresponding author or as otherwise indicated upon request. The individual genotype data for the FLCCA data are in dbGaP phs000716.v1.p1 (Genome-Wide Association Study of Lung Cancer Susceptibility in Never-Smoking Women in Asia). For the NCC and ACC studies, please contact Kouya Shiraishi at kshirais{at}ncc.go.jp or Takashi Kohno at tkkohno{at}ncc.go.jp for summary statistics. The GWAS data for the European populations contributing to this study are available at dbGap under accession phs000877.v1.p1 (Transdisciplinary Research Into Cancer of the Lung (TRICL), https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000876.v2.p1), phs001273.v3.p2 (Oncoarray Consortium, https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001273.v3.p2). To gain access to all data in dbGaP cited in this paper, please apply for dbGaP Authorized Access. GWAS data for the European populations contributing to this study are available under accession phs000877.v1.p1 (ILLCO), phs001273.v1.p1 (OncoArray).
Author contributions
B.B., N.R., Q.L. and H.Z. organized and designed the study. B.B., K.W., H.Z. carried out the staesecal analyses. B.B., K.W., Q.L. N.R. and H.Z. wrote the first dram. Q.L., M.T.L., B.A.B., W.H., N.E.C., BT.J., M.Song, H.P., D.A., C.C.C., L.B., M.Y., A.H., B.H., J.Liu, B.Zhu, S.I.B., C.K., K.WyaV, S.A.L., A.Chao, J.F.F.J., S.J.C., N.R., Z.Wang, C..L., J.C., C.W., W.T., D.Lin, SJ.A., XC.Z., J.S., YL.W., M.P.W., L.P.C., J.C.M.H., V.H.F.L., Z.H., K.M., J.Y..P., Jia.Liu, HS.J., J.E..C., Y.Y.C., H.N.K., MH.S., SS.K., YC.K., IJ.O., S.W.S., HI.Y., Y.T.K., YC.H., J.H.K., Y.H.K., J.S..S., Y.J..J., K.H.P., C.H.K., J.S.K., I.K.P., B..S., Jie.L, Z.W., S.C., J.Y., JC.W., Y.Y., YB.X., YT.G., D.L., J.Y.Y.W., H.C., L..J., J.Z., G.J., K..F., Z.Y., B.Z., W.W., P.G., Q.H., X.L., Y.R., A.S., Y.L., Y.C., WY.L., W.Z., XO.S., Q.C., G.Y., B.Q., T.W., H.G., L..L., P.X., F.W., G.W., J.X., J.L., R.CH.V., B.B., H.D.III.H., J.Wang, A.D.L..S., J.KC.C., V.L.S., K.C., H.Z., H.D., C.A.H., TY.C., LH.C…, IS.C., CY.C., S.S.J ., CH.C., GC.C., CF.H., YH.T., WC.W., KY.C., MS.H., WC.S., YM.C., CL.W., KC.C., CJ.Y., HH.H., FY.T.., HC.L., CJ.C., PC.Y., K.Shiraishi, T.K., H.K., S.M., H.H., K.Goto, Y.Ohe, S.W., Y.Yatabe, M.T., R.Hamamoto, A.Takahashi, Y.Momozawa, M.Kubo, Y.K., Y.D., Y.Miyagi, H.N., T.Y., N.S., M.I., M.H., Y.N., K.Takeuchi, K.W., K.Matsuda, Y.Murakami, K.S., K.T., Y.O., M.S., H.Suzuki, A.G., Y.M., T.H., M.K., K.O., H.S., J.D., H.M., M.Z., R.J.H., S.L., A.T., C.C., S.E.B., M.Johansson, A.R., H.Bö., HE.W., D.C., G.R., S.A., P.B., J.MK., J.K.F., S.S.S., L.L.M., O.M.., H.Bö., G.L., A.A., L.A.K., S.ZN., K.G., M.J., A.C., JM.Y., P.L., M.B.S., M.C.A., C.I.A., A.G.S., R.H., M.R.S., O.Y.G., I.P.G., X.W., P.Y. conducted epidemiology studies and contributed samples to GWAS and/or conducted inieal genotyping. All authors reviewed and approved the final review of the manuscript.
Competing interests
The authors declare no compeeng interests.
Abbreviations: SNP, single nucleotide polymorphism; OR, odds ratio; CI, confidence interval; AUC, area under the curve; PC, principle component; CT, clumping and thresholding
Data Availability
All data supporeng the findings described in this paper are available in the paper and in the Supplementary Informaeon and from the corresponding author or as otherwise indicated upon request. The individual genotype data for the FLCCA data are in dbGaP phs000716.v1.p1 (Genome-Wide Associaeon Study of Lung Cancer Suscepebility in Never-Smoking Women in Asia). For the NCC and ACC studies, please contact Kouya Shiraishi at kshirais{at}ncc.go.jp or Takashi Kohno at tkkohno{at}ncc.go.jp for summary staesecs. The GWAS data for the European populaeons contribueng to this study are available at dbGap under accession phs000877.v1.p1 (Transdisciplinary Research Into Cancer of the Lung (TRICL), https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000876.v2.p1), phs001273.v3.p2 (Oncoarray Consoreum, https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001273.v3.p2). To gain access to all data in dbGaP cited in this paper, please apply for dbGaP Authorized Access. GWAS data for the European populaeons contribueng to this study are available under accession phs000877.v1.p1 (ILLCO), phs001273.v1.p1 (OncoArray).
Code Availability
Methods implemented and their corresponding repositories include: SCT and LDpred2 at https://github.com/privefl/bigsnpr, PRS-CSx at https://github.com/geean107/PRScsx. CT-SLEB at https://github.com/andrewhaoyu/CTSLEB, and GENESIS at https://github.com/yandorazhang/GENESIS.
PLINK: https://www.cog-genomics.org/plink/1.9.
Most of our staesecal analyses were performed using the following R packages: ggplot2 v.3.3.3, dplyr v.1.0.4, data.table v.1.13.6, iCARE v.1.30.0 https://github.com/KevinWFred/PRS_EASLC
Never-Smoker Lung Cancer Working Group Baosen Zhou1, Jiucun Wang2,3, Wei Jie Seow4,5,6, Lei Song4, Wei Hu4, Li-Hsin Chien7,8, Qiuyin Cai9, Hee Nam Kim10, Yi-Long Wu11, Maria Pik Wong12, Brian Douglas Richardson4,13, Shilan Li4,14, Tongwu Zhang4, Charles Breeze4, Zhaoming Wang15, Bryan A Bassig16, Jin Hee Kim17, Demetrius Albanes4, Jason YY Wong18, Min-Ho Shin19, Lap Ping Chung12, Yang Yang20, She-Juan An11, Hong Zheng21, Yasushi Yatabe22, Xu-Chao Zhang11, Young-Chul Kim23,24, Neil E Caporaso4, Jiang Chang25, James Chung Man Ho26, Michiaki Kubo27, Yataro Daigo28,29, Minsun Song30, Yukihide Momozawa27, Yoichiro Kamatani31, Masashi Kobayashi32, Kenichi Okubo32, Takayuki Honda33, H Dean Hosgood34, Hideo Kunitoh35, Shun-ichi Watanabe36, Yohei Miyagi37, Haruhiko Nakayama38, Shingo Matsumoto39, Hidehito Horinouchi36, Masahiro Tsuboi40, Ryuji Hamamoto41, Koichi Goto39, Yuichiro Ohe36, Atsushi Takahashi31, Akiteru Goto42, Yoshihiro Minamiya43, Megumi Hara44, Yuichiro Nishida44, Kenji Takeuchi45, Kenji Wakai45, Koichi Matsuda46, Yoshinori Murakami47, Kimihiro Shimizu48, Hiroyuki Suzuki49, Motonobu Saito50, Yoichi Ohtaki51, Kazumi Tanaka51, Tangchun Wu52, Fusheng Wei53, Hongji Dai21, Mitchell J Machiela4, Jian Su11, Yeul Hong Kim54, In-Jae Oh23,24, Victor Ho Fun Lee55, Gee-Chen Chang56,57,58,59, Ying-Huang Tsai60,61, Kuan-Yu Che62, Ming-Shyan Huang63, Wu-Chou Su64, Yuh-Min Chen65, Adeline Seow5, Jae Yong Park66, Sun-Seog Kweon19,67, Kun-Chieh Chen57, Yu-Tang Gao68, Biyun Qian21, Chen Wu69, Daru Lu2,3, Jianjun Liu66,70,71, Ann G Schwartz72, Richard Houlston73, Margaret R Spitz74, Ivan P Gorlov74, Xifeng Wu75, Ping Yang76, Stephen Lam77, Adonina Tardon78, Chu Chen79, Seg E Bojesen80,81, Mawas Johansson82, Angela Risch83,84,85, Heike Bickeböller86, Bu-Tian Ji4, H-Erich Wichmann87,88,89, David C. Chriseani90, Gadi Rennert91, Susanne Arnold92, Paul Brennan82, James McKay82, John K Field93, Sanjay S Shete94, Loic Le Marchand95, Geoffrey Liu96, Angeline Andrew97, Lambertus A Kiemeney98, Shan Zienolddiny-Narui99, Kjell Grankvist100, Mikael Johansson101, Angela Cox102, Fiona Taylor102, Jian-Min Yuan103, Philip Lazarus104, MaVhew B Schabath105, Melinda C Aldrich106, Hyo-Sung Jeon107, Shih Sheng Jiang108, Jae Sook Sung54, Chung-Hsing Chen108, Chin-Fu Hsiao7, Yoo Jin Jung109, Huan Guo110, Zhibin Hu111, Laurie BurdeV4,112, Meredith Yeager4,112, Amy Hutchinson4,112, Belynda Hicks4,112, Jia Liu4,112, Bin Zhu4,112, Sonja I Berndt4, Wei Wu113, Junwen Wang114,115, Yuqing Li70, Jin Eun Choi107, Kyong Hwa Park54, Sook Whan Sung116, Li Liu117, Chang Hyun Kang109, Wen-Chang Wang118, Jun Xu119, Peng Guan113,120, Wen Tan69, Chong- Jen Yu121, Gong Yang9, Alan Dart Loon Sihoe122, Ying Chen5, Yi Young Choi107, Jun Suk Kim123, Ho- Il Yoon124, In Kyu Park109, Ping Xu125, Qincheng He113, Chih-Liang Wang126, Hsiao-Han Hung108,127, Roel C.H. Vermeulen128, Iona Cheng129, Junjie Wu2,3, Wei-Yen Lim5, Fang-Yu Tsai108,127, John K.C. Chan130, Jihua Li131, Hongyan Chen2,3, Hsien-Chih Lin7, Li Jin2,3, Jie Liu132, Norie Sawada133, Taiki Yamaji134, Kathleen WyaV4,112, Shengchao A. Li4,112, Hongxia Ma111,135, Meng Zhu111,135, Zhehai Wang132, Sensen Cheng132, Xuelian Li113,120, Yangwu Ren113,120, Ann Chao136, Motoki Iwasaki133,134, Junjie Zhu18, Gening Jiang18, Ke Fei18, Guoping Wu53, Chih-Yi Chen137, Chien-Jen Chen138, Pan-Chyr Yang139, Jinming Yu132, Victoria L. Stevens140, Joseph F. Fraumeni Jr4
1Department of Clinical Epidemiology and Center of Evidence Based Medicine, The First Hospital of China Medical University, Shenyang, China, 2Ministry of Educaeon Key Laboratory of Contemporary Anthropology, School of Life Sciences, Fudan University, Shanghai, China, 3State Key Laboratory of Geneec Engineering, School of Life Sciences, Fudan University, Shanghai, China, 4Division of Cancer Epidemiology and Geneecs, Naeonal Cancer Insetute, Rockville, MD, USA, 5Saw Swee Hock School of Public Health, Naeonal University of Singapore and Naeonal University Health System, Singapore, Singapore, 6Department of Medicine, Yong Loo Lin School of Medicine, Naeonal University of Singapore and Naeonal University Health System, Singapore, Singapore, 7Insetute of Populaeon Health Sciences, Naeonal Health Research Insetutes, Zhunan, Taiwan, 8Department of Applied Mathemaecs, Chung Yuan Chrisean University, Chong- Li, Taiwan, 9Division of Epidemiology, Department of Medicine, Vanderbilt University Medical Center and Vanderbilt-Ingram Cancer Center, Nashville, TN, USA, 10Department of Preveneve Medicine, Chonnam Naeonal University Medical School, Gwangju, Republic of Kore, 11Guangdong Lung Cancer Insetute, Guangdong Provincial Key Laboratory of Translaeonal Medicine in Lung Cancer, Guangdong Provincial People’s Hospital (Guangdong Academy of Medical Sciences), Southern Medical University, Guangzhou, China, 12Department of Pathology, Queen Mary Hospital, Hong Kong, Hong Kong, 13Department of Biostaesecs, Gillings School of Global Public Health, University of North Carolina, Chapel Hill, NC, USA, 14Department of Biostaesecs, Bioinformaecs & Biomathemaecs, Georgetown University Medical Center, Washington, DC, USA, 15Department of Computaeonal Biology, St. Jude Children’s Research Hospital, Memphis, TN, USA, 16Saville Cancer Screening and Preveneon Center, Inova Schar Cancer Insetute, Inova Health System, Fairfax, VA, USA, 17Department of Environmental Health, Graduate School of Public Health, Seoul Naeonal University, Seoul, Republic of Korea,, 18Epidemiology and Community Health Branch, Naeonal Heart Lung and Blood Insetute, Bethesda, MD, USA, 19Department of Preveneve Medicine, Seoul Naeonal University College of Medicine, Seoul, Republic of Korea, 20Shanghai Pulmonary Hospital, Shanghai, China, 21Department of Epidemiology and Biostaesecs, Naeonal Clinical Research Center for Cancer, Key Laboratory of Molecular Cancer Epidemiology of Tianjin, Tianjin Medical University Cancer Insetute and Hospital, Tianjin Medical University, Tianjin, China, 22Department of Pathology and Clinical Laboratories, Naeonal Cancer Center Hospital, Tokyo, Japan, 23Lung and Esophageal Cancer Clinic, Chonnam Naeonal University Hwasun Hospital, Hwasuneup, Republic of Korea, 24Department of Internal Medicine, Chonnam Naeonal University Medical School, Gwangju, Republic of Korea, 25Department of Eeology & Carcinogenesis, Cancer Insetute and Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 26Department of Medicine, School of Clinical Medicine, The University of Hong Kong, Queen Mary Hospital, Hong Kong, Hong Kong, 27Laboratory for Genotyping Development, RIKEN Center for Integraeve Medical Sciences, Yokohama, Japan, 28Center for Anebody and Vaccine Therapy, Research Hospital, Insetute of Medical Science, The University of Tokyo, Tokyo, Japan, 29Department of Medical Oncology and Cancer Center, and Center for Advanced Medicine against Cancer, Shiga University of Medical Science, Shiga, Japan, 30Department of Staesecs & Research Insetute of Natural Sciences, Sookmyung Women’s University, Seoul, 04310, Korea, 31Laboratory for Staesecal Analysis, RIKEN Center for Integraeve Medical Sciences, Yokohama, Japan, 32Department of Thoracic Surgery, Tokyo Medical and Dental University, Tokyo, Japan, 33Department of Respiratory Medicine, Tokyo Medical and Dental University, Tokyo, Japan, 34Department of Epidemiology and Populaeon Health, Albert Einstein College of Medicine, New York, USA, 35Department of Medical Oncology, Japanese Red Cross Medical Center, Tokyo, Japan, 36Department of Thoracic Surgery, Naeonal Cancer Center Hospital, Tokyo, Japan, 37Molecular Pathology and Geneecs Division, Kanagawa Cancer Center Research Insetute, Yokohama, Japan, 38Department of Thoracic Surgery, Kanagawa Cancer Center, Yokohama, Japan, 39Department of Thoracic Oncology, Naeonal Cancer Center Hospital East, Kashiwa, Japan, 40Department of Thoracic Surgery, Naeonal Cancer Center Hospital East, Kashiwa, Japan, 41Division of Medical AI Research and Development, Naeonal Cancer Center Research Insetute, Tokyo, Japan, 42Department of Cellular and Organ Pathology, Graduate School of Medicine, Akita University, Akita, Japan, 43Department of Thoracic Surgery, Graduate School of Medicine, Akita University, Akita, Japan, 44Department of Preveneve Medicine, Faculty of Medicine, Saga University, Saga, Japan, 45Department of Preveneve Medicine, Nagoya University Graduate School of Medicine, Nagoya, Japan, 46Laboratory of Clinical Genome Sequencing, Department of Computaeonal Biology and Medical Science, Graduate School of Froneer Sciences, The University of Tokyo, Tokyo, Japan, 47Department of Molecular Biology, Insetute for Advanced Medical Sciences, Nippon Medical School, Tokyo, Japan, 48Department of Surgery, Division of General Thoracic Surgery, Shinshu University School of Medicine Asahi, Nagano, Japan, 49Department of Chest Surgery, Fukushima Medical University School of Medicine, Fukushima, Japan, 50Department of Gastrointesenal Tract Surgery, Fukushima Medical University School of Medicine, Fukushima, Japan, 51Department of Integraeve center of General Surgery, Gunma University Hospital, Gunma, Japan, 52Insetute of Occupaeonal Medicine and Ministry of Educaeon Key Lab for Environment and Health, School of Public Health, Huazhong University of Science and Technology, Wuhan, China, 53China Naeonal Environmental Monitoring Center, Beijing, China, 54Department of Internal Medicine, Division of Oncology/Hematology, College of Medicine, Korea University Anam Hospital, Seoul, Republic of Korea, 55Department of Clinical Oncology, The University of Hong Kong, Queen Mary Hospital, Hong Kong, Hong Kong, 56School of Medicine and Insetute of Medicine, Chung Shan Medical University, Taichung, Taiwan, 57Department of Internal Medicine, Division of Pulmonary Medicine, Chung Shan Medical University Hospital, Taichung, Taiwan, 58Insetute of Biomedical Sciences, Naeonal Chung Hsing University, Taichung, Taiwan, 59Department of Internal Medicine, Division of Chest Medicine, Taichung Veterans General Hospital, Taichung, Taiwan, 60Department of Respiratory Therapy, Chang Gung University, Taoyuan, Taiwan, 61Department of Pulmonary and Criecal Care, Xiamen Chang Gung Hospital, Xiamen, China, 62Department of Internal Medicine, Naeonal Taiwan University Hospital and College of Medicine, Taipei, Taiwan, 63Department of Internal Medicine, E-Da Cancer Hospital, I-Shou University and Kaohsiung Medical University, Kaohsiung, Taiwan, 64Department of Oncology, Naeonal Cheng Kung University Hospital, College of Medicine, Naeonal Cheng Kung University, Tainan, Taiwan, 65Department of Chest Medicine, Taipei Veterans General Hospital, and school of Medicine, Naeonal Yang Ming Chiao Tung University, Taipei, Taiwan, 66Lung Cancer Center, Kyungpook Naeonal University Medical Center, Daegu, Republic of Korea, 67Jeonnam Regional Cancer Center, Chonnam Naeonal University, Hwasun, Republic of Korea, 68Department of Epidemiology, Shanghai Cancer Insetute, Shanghai, China, 69Department of Eeology & Carcinogenesis and State Key Laboratory of Molecular Oncology, Cancer Insetute and Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China, 70Department of Human Geneecs, Genome Insetute of Singapore, Singapore, Singapore, 71School of Life Sciences, Anhui Medical University, Hefei, China, 72Karmanos Cancer Insetute, Wayne State University, Detroit, MI, USA, 73Division of Geneecs and Epidemiology, Insetute of Cancer Research, London, UK, 74Department of Medicine, Seceon of Epidemiology and Populaeon Science, Insetute for Clinical and Translaeonal Research, Houston, TX, USA, 75School of Medicine, Zhejiang University, Hangzhou, Zhejiang, China, 76Department of Health Sciences Research, Mayo Clinic, ScoVsdale, AZ, USA, 77Department of Integraeve Oncology, BC Cancer, Vancouver, BC, Canada, 78IUOPA, University of Oviedo and CIBERESP, Spain, 79Program in Epidemiology, Public Health Sciences Division, Fred Hutchinson Cancer Center, SeaVle, Washington, USA, 80Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark, 81Department of Clinical Biochemistry, Herlev and Gentome Hospital, Copenhagen University Hospital, Copenhagen, Denmark, 82Internaeonal Agency for Research on Cancer (IARC/WHO), Lyon, France, 83German Cancer Research Center (DKFZ), Heidelberg, Germany, 84Translaeonal Lung Research Center Heidelberg (TLRC-H), Member of the German Center for Lung Research (DZL), Heidelberg, Germany, 85University of Salzburg and Cancer Cluster Salzburg, Salzburg, Austria, 86University Medical Center Goewngen, Goewngen, Germany, 87Insetute of Medical Informaecs, Biometry and Epidemiology, Ludwig Maximilians University, Munich, Germany, 88Helmholtz Center Munich, Insetute of Epidemiology, Munich, Germany, 89Insetute of Medical Staesecs and Epidemiology, Technical University Munich, Munich, Germany, 90Harvard TH Chan School of Public Health, Boston, MassachuseVs, USA, 91Carmel Medical Center, Israel, 92Markey Cancer Center, Lexington, KY, USA, 93Liverpool University, Liverpool, UK, 94The University of Texas MD Anderson Cancer Center, Houston, Texas, USA, 95Epidemiology Program, University of Hawaii Cancer Center, Honolulu, HI, USA, 96Princess Margaret Cancer Center, Toronto, ON, Canada, 97Norris CoVon Cancer Center, Lebanon, NH, USA, 98Radboud University Medical Center, Nijmegen, Netherlands, 99Naeonal Insetute of Occupaeonal Health, Oslo, Norway, 100Department of Medical Biosciences, Umeå University, Umeå, Sweden, 101Department of Radiaeon Sciences, Umeå University, Umeå, Sweden, 102University of Sheffield, Sheffield, UK, 103UPMC Hillman Cancer Center and Department of Epidemiology, School of Public Health, University of PiVsburgh, PiVsburgh, PA, USA, 104Department of Pharmaceuecal Sciences, College of Pharmacy and Pharmaceuecal Sciences, Washington State University, Spokane, WA, USA, 105Department of Cancer Epidemiology, H. Lee MoffiV Cancer Center and Research Insetute, Tampa, FL, USA, 106Department of Thoracic Surgery, Division of Epidemiology, Vanderbilt University Medical Center, Nashville, TN, USA, 107Cancer Research Center, Kyungpook Naeonal University Medical Center, Daegu, Republic of Korea, 108Naeonal Insetute of Cancer Research, Naeonal Health Research Insetutes, Zhunan, Taiwan, 109Department of Thoracic and Cardiovascular Surgery, Cancer Research Insetute, Seoul Naeonal University College of Medicine, Seoul, Republic of Korea, 110Department of Occupaeonal and Environmental Health and Ministry of Educaeon Key Lab for Environment and Health, School of Public Health, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China, 111Department of Epidemiology, School of Public Health, Nanjing Medical University, Nanjing, China, 112Cancer Genomics Research Laboratory, Frederick Naeonal Laboratory for Cancer Research, Leidos Biomedical Research Inc, Rockville, MD, USA, 113Department of Epidemiology, School of Public Health, China Medical University, Shenyang, China, 114Division of Applied Oral Sciences & Community Dental Care, Faculty of Denestry, The University of Hong Kong, Hong Kong, China, 115State Key LaState Key Laboratory of Pharmaceuecal Biotechnology, The University of Hong Kong, Hong Kong, Chinaboratory of Pharmaceuecal Biotechnology, The University of Hong Kong, Hong Kong, China, 116State Key Laboratory of Pharmaceuecal Biotechnology, The University of Hong Kong, Hong Kong, China, 117Department of Oncology, Cancer Center, Union Hospital, Huazhong University of Science and Technology, Wuhan, China, 118The Ph.D. Program for Translaeonal Medicine, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan, 119School of Public Health, Li Ka Shing (LKS) Faculty of Medicine, The University of Hong Kong, Hong Kong, China, 120Key Laboratory of Cancer Eeology and Interveneon, University of Liaoning Province, Shenyang, China, 121Department of Internal Medicine, Naeonal Taiwan University Hospital Hsin-Chu Branch, Hsinchu, Taiwan, 122Gleneagles Hong Kong Hospital, Hong Kong, China, 123Department of Internal Medicine, Division of Medical Oncology, College of Medicine, Korea University Guro Hospital, Seoul, Republic of Korea, 124Department of Internal Medicine, Seoul Naeonal University Bundang Hospital, Seongnam, Republic of Korea, 125Department of Oncology, Wuhan Iron and Steel (Group) Corporaeon Staff-Worker Hospital, Wuhan, China, 126Division of Pulmonary Oncology and Interveneonal Bronchoscopy, Department of Thoracic Medicine, Chang Gung Memorial Hospital, Taoyuan, Taiwan, 127Division of Thoracic Surgery, Department of Surgery, Chung Shan Medical University Hospital, Taichung, Taiwan, 128Division of Environmental Epidemiology, Insetute for Risk Assessment Sciences (IRAS), Utrecht University, Utrecht, The Netherlands, 129Department of Epidemiology and Biostaesecs, University of California, San Francisco, San Francisco, CA, USA, 130Department of Pathology, Queen Elizabeth Hospital, Hong Kong, China, 131Qujing Center for Diseases Control and Preveneon, Qujing, China, 132Department of Oncology, Shandong Cancer Hospital and Insetute, Shandong Academy of Medical Sciences, Jinan, China, 133Division of Cohort Research, Naeonal Cancer Center Insetute for Cancer Control, Tokyo, Japan, 134Division of Epidemiology, Naeonal Cancer Center Insetute for Cancer Control, Tokyo, Japan, 135Jiangsu Key Lab of Cancer Biomarkers, Preveneon and Treatment, Collaboraeve Innovaeon Center for Cancer Medicine, Nanjing Medical University, Nanjing, China, 136Center for Global Health, Naeonal Cancer Insetute, Bethesda, MD, USA, 137Insetute of Medicine, Chung Shan Medical University, Taichung, Taiwan, 138Genomic Research Center, Academia Sinica, Taipei, Taiwan, 139Department of Internal Medicine, Naeonal Taiwan University Hospital, Taipei, Taiwan, 140Laboratory Services, American Cancer Society, Georgia, USA
Acknowledgements
This work uelized the computaeonal resources of the high-performance computaeon Biowulf cluster at Naeonal Insetutes of Health, USA (hVp://hpc.nih.gov).
Female Lung Cancer Consoreum in Asia (NCI): This study was supported by a Grant-in-Aid for Scienefic Research on Priority Areas from the Ministry of Educaeon, Science, Sports, Culture and Technology of Japan, a Grant-in- Aid for the Third Term Comprehensive 10-Year Strategy for Cancer Control from the Ministry Health, Labor and Welfare of Japan, by Health and Labor Sciences Research Grants for Research on Applying Health Technology from the Ministry of Health, Labor and Welfare of Japan, by the Naeonal Cancer Center Research and Development Fund, the Naeonal Research Foundaeon of Korea (NRF) grant funded by the Korea government (MEST) (grant No. 2011-0016106), a grant of the Naeonal Project for Personalized Genomic Medicine, Ministry for Health & Welfare, Republic of Korea (A111218-11-GM04), the Program for Changjiang Scholars and Innovaeve Research Team in University in China (IRT_14R40 to K.C.), the Naeonal Science & Technology Pillar Program (2011BAI09B00), MOE 111 Project (B13016), the Naeonal Natural Science Foundaeon of China (No. 30772531, and 81272618), Guangdong Provincial Key Laboratory of Lung Cancer Translaeonal Medicine (No. 2012A061400006), Special Fund for Research in the Public Interest from the Naeonal Health and Family Planning Commission of PRC (No. 201402031), and the Ministry of Science and Technology, Taiwan (MOST 103-2325-B-400-023 & 104-2325-B-400-012). The Japan Lung Cancer Study (JLCS) was supported in part by the Pracecal Research for Innovaeve Cancer Control from Japan Agency for Medical Research and Development (15ck0106096h0002) and the Management Expenses Grants from the Government to the Naeonal Cancer Center (26-A-1) for Biobank. BioBank Japan was supported by the Ministry of Educaeon, Culture, Sports, Sciences and Technology of the Japanese government. The Japan Public Health Center-based prospeceve Study (the JPHC Study) was supported by the Naeonal Cancer Center Research and Development Fund (23-A- 31[toku], 26-A-2, 29-A-4, 2020-J-4, and 2023-J-4) (since 2011) and a Grant-in-Aid for Cancer Research from the Ministry of Health, Labour and Welfare of Japan (from 1989 to 2010). The Taiwan GELAC Study (Geneec Epidemiological Study for Lung AdenoCarcinoma) was sup- ported by grants from the Naeonal Research Program on Genomic
Medicine in Taiwan (DOH99-TD-G-111-028), the Naeonal Research Program for Biopharmaceuecals in Taiwan (MOHW 103-TDUPB-211-144003, MOST 103-2325-B-400-023) and the Bioinformaecs Core Facility for Translaeonal Medicine and Biotechnology Development (MOST 104-2319-B-400-002). This work was also supported by the Jinan Science Research Project Foundaeon (201102051), the Naeonal Key Scienefic and Technological Project (2011ZX09307-001-04), the Naeonal Natural Science Foundaeon of China (No.81272293), the State Key Program of Naeonal Natural Science of China (81230067), the Naeonal Research Foundaeon of Korea (NRF) grant funded by the Korea government (MSIP) (No. NRF- 2014R1A2A2A05003665), Sookmyung Women’s University Research Grants, Korea (1-1603- 2048), Agency for Science, Technology and Research (A*STAR), Singapore and the US Naeonal Insetute of Health Grant (1U19CA148127-01). The overall GWAS project was supported by the intramural program of the US Naeonal Insetutes of Health/Naeonal Cancer Insetute. The following is a list of grants by study center: SKLCS (Y.T.K.)—Naeonal Research Foundaeon of Korea (NRF) grant funded by the Korea government (MEST) (2011-0016106). (J.C.) – This work was supported by a grant from the Naeonal R&D Program for Cancer Control, Ministry of Health &Welfare, Republic of Korea (grant no. 0720550-2). (J.S.S) – grant number is A010250. WLCS (T.W.)—Naeonal Key Basic Research and Development Program (2011CB503800). SLCS (B.Z.)— Naeonal Nature Science Foundaeon of China (81102194). Liaoning Provincial Department of Educaeon (LS2010168). China Medical Board (00726). GDS (Y.L.W.)—Foundaeon of Guangdong Science and Technology Department (2006B60101010, 2007A032000002, 2011A030400010).
Guangzhou Science and Informaeon Technology Bureau (2011Y2-00014). Chinese Lung Cancer Research Foundaeon, Naeonal Natural Science Foundaeon of China (81101549). Natural Science Foundaeon of Guangdong Province (S2011010000792). TLCS (K.C., B.Q)—Program for Changjiang Scholars and Innovaeve Research Team in University (PCSIRT), China (IRT1076).
Tianjin Cancer Insetute and Hospital. Naeonal Foundaeon for Cancer Research (US). FLCS (J.C.W., D.R., L.J.)—Ministry of Health (201002007). Ministry of Science and Technology (2011BAI09B00). Naeonal S&T Major Special Project (2011ZX09102-010-01). China Naeonal High-Tech Research and Development Program (2012AA02A517, 2012AA02A518). Naeonal Science Foundaeon of China (30890034). Naeonal Basic Research Program (2012CB944600).
Scienefic and Technological Support Plans from Jiangsu Province (BE2010715). NLCS (H.S.)— China Naeonal High-Tech Research and Development Program Grant (2009AA022705). Priority Academic Program Development of Jiangsu Higher Educaeon Insetueon. Naeonal Key Basic Research Program Grant (2011CB503805). GEL-S (A.S.)—Naeonal Medical Research Council Singapore grant (NMRC/0897/2004, NMRC/1075/2006). (J.Liu)—Agency for Science, Technology and Research (A*STAR) of Singapore. GELAC (C.A.H.)—Naeonal Research Program on Genomic Medicine in Taiwan (DOH98-TDG-111-015). Naeonal Research Program for Biopharmaceuecals in Taiwan (DOH 100- TD-PB-111-TM013). Naeonal Science Council, Taiwan (NSC 100- 2319-B- 400-001). YLCS (Q.L.)—Supported by the intramural pro- gram of U.S. Naeonal Insetutes of Health, Naeonal Cancer Insetute. SWHS (W.Z., W-H.C., N.R.)—The work was supported by a grant from the Naeonal Insetutes of Health (R37 CA70867, UM1 CA182910) and the Naeonal Cancer Insetute intramural research program, including NCI Intramural Research Program contract (N02 CP1101066). JLCS (K.M., T.K.)—Grants-in-Aid from the Ministry of Health, Labor, and Welfare for Research on Applying Health Technology and for the 3rd-term Comprehensive 10-year Strategy for Cancer Control; by the Naeonal Cancer Center Research and Development Fund; by Grant-in-Aid for Scienefic Research on Priority Areas and on Innovaeve Area from the Ministry of Educaeon, Science, Sports, Culture and — Technology of Japan. (W.P.)—NCI R01- CA121210. HKS (J.W.)— General Research Fund of Research Grant Council, Hong Kong (781511M). The Environment and Geneecs in Lung Cancer Eeology (EAGLE), Prostate, Lung, Colon, Ovary Screening Trial (PLCO), and Alpha-Tocopherol, Beta-Carotene Cancer Preveneon (ATBC) studies were supported by the Intramural Research Program of the Naeonal Insetutes of Health, Naeonal Cancer Insetute (NCI), Division of Cancer Epidemiology and Geneecs. ATBC was also supported by U.S. Public Health Service contracts (N01-CN-45165, N01-RC-45035, and N01- RC-37004) from the NCI. PLCO was also supported by individual contracts from the NCI to the University of Colorado Denver (NO1-CN-25514), Georgetown University (NO1-CN-25522), the Pacific Health Research Insetute (NO1-CN-25515), the Henry Ford Health System (NO1-CN- 25512), the University of Minnesota, (NO1-CN- 25513), Washington University (NO1-CN-25516), the University of PiVsburgh (NO1-CN-25511), the University of Utah (NO1-CN- 25524), the Marshfield Clinic Research Foundaeon (NO1-CN- 25518), the University of Alabama at Birmingham (NO1-CN- 75022), Westat, Inc. (NO1-CN-25476), and the University of California, Los Angeles (NO1-CN-25404). The Carotene and Reenol Efficacy Trial (CARET) is funded by the Naeonal Cancer Insetute, Naeonal Insetutes of Health through grants U01-CA063673, UM1- CA167462, and U01-CA167462. The Cancer Preveneon Study-II (CPS-II) Nutrieon Cohort was supported by the American Cancer Society. The NIH Genes, Environment and Health Inieaeve (GEI) partly funded DNA extraceon and staes- ecal analyses (HG-06-033-NCI-01 and RO1HL091172-01), genotyping at the Johns Hopkins University Center for Inherited Disease Research. This research was supported by the Naeonal Research Foundaeon of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2020R1A2C4002236) Female Lung Cancer Consoreum in Asia (Tianjin): Tianjin Science and Technology CommiVee Foundaeon, 18YFZCSY00520.
Female Lung Cancer Consoreum in Asia (Taiwan): The Ministry of Health and Welfare grants DOH97-TD-G-111-028 (ISC), DOH98-TD-G-111-017 (ISC), DOH99-TD-G-111-014 (ISC); DOH97- TD-G-111-026 (CAH), DOH98-TD-G-111-015 (CAH), DOH99-TD-G-111-028 (CAH); Naeonal Health Research Insetutes grants NHRI-PH-110-GP-01, NHRI-PH-110-GP-03; and the Ministry of Science and Technology grants MOST108-2314-B-400-038(CAH), MOST109-2740-B-400- 002(CAH), MOST 111-2740-B-400-002 (CAH), MOST 111-2314-B-400-020 (CAH).
The GWAS of lung cancer in European never smokers was supported by NIH R01 CA149462 (OYG).
OncoArray study in Europeans: The OncoArray data and analysis from INTEGRAL-ILCCO were supported by NIH U19 CA203654, and U19 CA148127. The data harmonizaeon for ILCCO was supported by Canadian Insetute for Health Research (CIHR) Canada Research Chair to R.J.H, and CIHR FDN 167273).
European never-smoking lung cancer study: CIA is a Research Scholar of the Cancer Preveneon Insetute of Texas (CPRIT) and supported by CPRIT grant RR170048.
Taiwan eQTL study: This study was supported by the Ministry of Health and Welfare grants DOH97-TD-G-111-028 (ISC), DOH98-TD-G-111-017 (ISC), DOH99-TD-G-111-014 (ISC); DOH97- TD-G-111-026 (CAH), DOH98-TD-G-111-015 (CAH), DOH99-TD-G-111-028 (CAH); Naeonal Health Research Insetutes grants NHRI-PH-110-GP-01, NHRI-PH-110-GP-03; and the Ministry of Science and Technology grants MOST108-2314-B-400-038(CAH), MOST109-2740-B-400- 002(CAH).
B.B., K.W., J.C., J.S. N.R., Q.L. and H.Z. are supported by NIH intramural Research Program. N.C. is supported by NIH grant 1R01HG010480. P.Y. is supported by Mayo Clinic Foundaeon Research Funds, NIH-CA77118 and CA80127. G.L. is supported is supported by the Alan Brown Chair and Lusi Wong Fund of the Princess Margaret Cancer Foundaeon. D.C.C is supported by U01CA209414. O.Y.G. is supported by NIH R01 CA231141.
Footnotes
↵‡ These authors jointly directed this work
References
