Supplementary Table S7. Confounding analyses for the association between the interaction of the TGF-B-Polygenic Risk Score and red meat intake in relation to colorectal cancer risk.

BACKGROUND:Integrating genetic and lifestyle information has the potential to greatly improve the prediction of colorectal cancer (CRC) risk. However, racial and ethnic minorities are generally underrepresented in gene-environment studies of CRC risk. METHODS:We investigated the interplay of genetics and lifestyle on CRC risk in a prospective analysis of 68 397 African American, Japanese American, Latino, Native Hawaiian, and White individuals from the Multiethnic Cohort Study. Genetic predisposition was assessed using a 205-variant polygenic risk score. Lifestyle was assessed using a lifestyle risk score based on smoking, alcohol consumption, body mass index, physical activity, and diet. The independent and joint associations of the polygenic risk score and lifestyle risk score on CRC risk were evaluated using Cox regression. RESULTS:We identified 1303 incident CRC cases (median 15.1-year follow-up). The highest quintile of the polygenic risk score was associated with a 2.4-fold increase in CRC risk compared with the lowest quintile (Q5 vs Q1: hazard ratio [HR] = 2.40, 95% confidence interval [CI] = 1.99 to 2.89). The highest quintile of the lifestyle risk score was associated with a 54% increased risk (Q5 vs Q1: HR = 1.54, 95% CI = 1.26 to 1.88). This association was stronger among those with high genetic risk (polygenic risk score  ≥ 50%; Q5 vs Q1: HR = 1.82, 95% CI = 1.41 to 2.35) and statistically nonsignificant among those with low genetic risk (polygenic risk score  < 50%; Q5 vs Q1: HR = 1.20, 95% CI = 0.88 to 1.64; P for interaction = 0.01). Results were similar across race and ethnicity. CONCLUSIONS:Our study suggests that lifestyle modification may offer greater risk reduction among those at higher genetic risk. Future research is warranted to enhance the integration of genetics and lifestyle in CRC risk stratification and screening approaches across populations.

AbstractBackground: Healthy dietary patterns have been linked to reduced cancer risk, but evidence for prostate cancer remains inconsistent, especially across diverse populations. Methods: In the Multiethnic Cohort study, we examined associations between 11 diet scores and prostate cancer risk among 79,930 men (White, African American, Japanese American, Latino, or Native Hawaiian). HRs and 95% confidence intervals (CI) per 1 SD increase in each score were estimated from Cox proportional hazards models for total prostate cancer and by grade, stage, and aggressiveness. Analyses were conducted overall and within racial/ethnic groups. A significance threshold of P < 0.003 was used to adjust for multiple testing. Results: Over a mean follow-up of 18.8 years, 9,759 prostate cancer cases were identified. No significant associations were observed in the hypothesized direction between dietary scores and prostate cancer risk in the overall population. However, among African Americans, higher Dietary Approaches to Stop Hypertension scores were suggestively associated with a 14% lower risk of advanced prostate cancer (95% CI, 0.77–0.98; P = 0.02). Among Japanese Americans, higher Empirical Dietary Index for Insulin Resistance scores were linked to an increased risk of low-grade (HR = 1.09; 95% CI, 1.02–1.17; P = 0.01) and nonaggressive prostate cancer (HR = 1.10; 95% CI, 1.02–1.19; P = 0.02). Conclusions: Although no strong associations were found overall, specific dietary patterns may influence prostate cancer risk differently by race/ethnicity and tumor subtype, warranting further investigation in large and diverse populations. Impact: These findings highlight the need for future studies to explore disease-specific and culturally informed dietary patterns that may inform prostate cancer prevention across diverse populations.

Supplementary Figure S1, LocusZoom plots.

Abstract Inherited susceptibility plays a critical role in prostate cancer (PCa) risk. Using large biobank and case-control datasets, we evaluated the contribution of both common variants and rare germline pathogenic variants (PVs) to overall PCa risk. The Prostate Cancer Exome Sequencing Consortium currently includes 427,388 male participants (51,452 PCa cases and 375,936 controls) with whole-exome sequencing data from ten biobanks and studies: UK Biobank (14,669 cases/195,600 controls), All of Us Research Program (7,577/75,226), African Ancestry Prostate Cancer Consortium (7,176/4,675), Mayo Clinic Biobank (6,031/15,084), Mass General Brigham Biobank (3,393/14,095), Geisinger’s MyCode Community Health Initiative (3,026/15,130), UCLA ATLAS Precision Medicine Biobank (2,850/16,904), Penn Medicine Biobank (2,598/16,255), Colorado Center for Personalized Medicine (2,269/13,399), and Malmo Diet and Cancer (1,863/9,568). Based on self-reported race/ethnicity and estimated genetic ancestry, the cases comprise approximately 79% European, 18% African, and 3% other ancestry populations. Single-variant association analyses tested all variants on chromosomes 1-22 and X with a minor allele count ≥ 5. In gene-based analyses, PVs were defined as rare variants (minor allele frequency [MAF] < 1% in controls) that had either a Variant Effect Predictor (VEP) impact score of “high” or a pathogenic or likely pathogenic ClinVar classification. Associations were estimated using Firth logistic regression, adjusting for age and the top ten genetic principal components. Results from individual studies were combined using fixed-effect meta-analysis. In single-variant association analyses, 496 variants reached genome-wide significance (p<5×10-8; MAF>0.02%). Among these, 458 (92%) variants mapped to previously known risk regions, including three rare PVs in HOXB13 (rs138213197), CHEK2 (rs555607708), and FAM111A (rs533676902). Characterization of the remaining 38 variants is ongoing. Gene-based analyses identified significant associations (p<2.4×10-6) for eight genes: HOXB13 (OR=3.7, 95% CI=3.3-4.2), BRCA2 (OR=2.0, 95% CI=1.7-2.3), CHEK2 (OR=1.6, 95% CI=1.5-1.8), ATM (OR=1.6, 95% CI=1.4-1.9), FAM111A (OR=1.4, 95% CI=1.3-1.5), BIK (OR=1.4, 95% CI=1.2-1.6), SAMHD1 (OR=2.1, 95% CI=1.6-2.7), and SMOC2 (OR=3.2, 95% CI=2.0-5.1). All genes except SMOC2 have been previously implicated in PCa susceptibility. Among cancer predisposition and DNA repair genes, nominal associations were also observed for XRCC2 (OR=1.6, 95% CIs=1.2-2.3) and BRCA1 (OR=1.2, 95% CI=1.0-1.4), whereas the association was not significant for PALB2 (OR=1.2, 95% CI=0.9-1.5). These findings reinforce the role of rare germline PVs, particularly in cancer predisposition and DNA repair genes, in PCa susceptibility. As additional studies are incorporated into the Consortium, we expect this work to provide a more comprehensive characterization of the genetic architecture of PCa. Citation Format: Yifan Zhang, Shuyan Cheng, Nicholas Boddicker, Matthew Lebo, Alexander S. Berry, Roni Haas, Ryan Hausler, Tokhir Dadaev, Heena Desai, Alex A. Rodriguez, Ravi K. Madduri, Andrew Hill, Xin Sheng, Susan M. Gundell, Mine Cicek, Penn Medicine Biobank, Olle Melander, Chris R. Gignoux, Isla P. Garraway, Bogdan Pasaniuc, Paul C. Boutros, Matt Oetjens, Adam S. Kibel, Robert J. Klein, Zsofia Kote-Jarai, Fergus J. Couch, Kara N. Maxwell, Burcu F. Darst, David V. Conti, Christopher A. Haiman, Fei Chen. Genetic risk of prostate cancer: Insights from the Prostate Cancer Sequencing Consortium [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr LB390.

Associations of dietary scores (per 1-SD increase) with prostate cancer risk among Japanese Americans in the MEC.

Supplementary Table S3. Weights used for polygenic risk score estimation and coordinates from single nucleotide polymorphisms considered in the study.

Motivation High-dimensional omics data are typically measured on limited sample sizes, which challenges model-based clustering methods such as Gaussian mixture models (GMMs), often leading to instability and poor generalization under complex mixture structures. To address these limitations, we developed Praxis-BGM, a natural-gradient variational inference framework for GMMs. Praxis-BGM enables semi-supervised transfer learning by incorporating an informative prior GMM estimated from large-scale reference data with robust cluster structures. The prior model can encode cluster-specific means, covariance structures, and structural connectivity patterns, and is updated using the target data with variational inference to improve clustering in small-sample settings.Results Using the Variational Online Newton (VON) algorithm, we derived natural-gradient updates for the standard parameters of GMMs. Implemented in the Python library JAX for accelerator-oriented computation, Praxis-BGM is computationally efficient and scalable. Across extensive simulations and two real-world applications-breast cancer bulk transcriptomics for subtype recovery and single-cell transcriptomics for cross-platform cell-type label transfer-Praxis-BGM improves posterior clustering performance, stability, and biological interpretability, even when priors are partially mismatched.Availability and implementation Praxis-BGM is freely available at https://github.com/ContiLab-usc/Praxis-BGM, and an archival version is available on Zenodo at https://doi.org/10.5281/zenodo.19657680.

Associations of dietary scores (per 1-SD increase) with prostate cancer risk among Whites in the MEC.

Abstract Background: Prostate cancer (PCa) is the second most frequently diagnosed malignancy among men. While multiple protein markers have been implicated in PCa, findings from conventional studies are often inconsistent due to methodological limitations such as selection bias and uncontrolled confounding. The proteome-wide association study (PWAS) design leverages genetic instruments to identify protein biomarkers with potential causal roles in diseases. Although candidate causal proteins in blood have been identified for PCa in our previous work, few studies have focused on prostate tissue. Methods: We conducted the first prostate tissue-based PWAS using data from the PRACTICAL/ELLIPSE consortia, comprising 122,188 PCa cases and 604,640 controls. Proteomic and genomic data were generated from 201 frozen prostate tissue samples without PCa, quantifying 11,575 proteins. We used data of 195 unrelated subjects for model building. Prediction models for protein abundance were built using nearby unambiguous SNPs of potentially associated variants, applying BLUP, LASSO, elastic net, and top1 methods. Association testing was performed for genetically predicted protein levels with PCa risk and aggressiveness. For one of the identified proteins eIF4G1, we performed knockdown of its gene expression in androgen-sensitive (LNCaP), enzalutamide-resistant LNCaP (NO-LNCaP-ENZR), and castration-resistant (22RV1) PCa cell lines, and investigated the effects on multiple phenotypes. Survival analysis was also performed using TCGA primary PCa RNA-seq data. Results: A total of 1,034 protein models achieved cross-validated R2 > 0.01 and were retained for association testing. Fifty-six proteins showed significant associations with PCa risk, including 18 associated with advanced disease and seven distinguishing advanced from non-advanced cases. One of the top novel proteins, EIF4G1, is required for the initial steps of translation. Disrupting eIF4F complex activity via EIF4G1 knockdown reduced cell proliferation, colony formation, and spheroid culture growth, and decreased cell migration and invasion. EIF4G1 knockdown also sensitized LNCaP cells to enzalutamide treatment and inhibited clonogenic potential of enzalutamide-resistant cells. Pharmacological inhibition of eIF4F with SBI-756 reproduced these effects and induced G1 phase cell cycle arrest. Polysome profiling revealed decreased mRNA loading onto polysomes, indicating that knockdown of EIF4G1 impaired cap-dependent translation. Elevated expression of EIF4G1 in tumors was also associated with shorter disease-specific survival. Conclusions: Our study reveals novel prostate tissue proteins putatively causally linked to PCa risk and aggressiveness. Our functional work suggests that a novel protein, eIF4G1, presents a new target for limiting PCa progression and overcoming therapy resistance. Citation Format: Jingjing Zhu, Pramod KC, Sweaty Koul, Yijun Tian, Hua Zhong, Thomas G. Beach, Hyeyoon Kim, Athena A. Schepmoes, Karl K. Weitz, Tyler Sagendorf, Tao Liu, Maarit I. Tiirikainen, Lucio Miele, Nicholas Mancuso, Timothy R. Rebbeck, David V. Conti, Christopher A. Haiman, the PRACTICAL/ELLIPSE consortium, Chong Wu, Liang Wang, Hari K. Koul, Lang Wu. Uncovering causal protein markers in prostate tissue for prostate cancer: A proteome-wide association study and functional validation [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6322.

Abstract Background: Adherence to a healthy lifestyle has been consistently associated with reduced breast cancer risk, but whether this benefit varies by genetic susceptibility remains unclear, particularly among racially and ethnically diverse populations. This study evaluated whether the association between a Healthy Lifestyle Index Score (HLIS) and breast cancer risk differs by a 313-variant polygenic risk score (PRS) among postmenopausal women in the Multiethnic Cohort (MEC). Methods: HLIS (range: 0-7) was constructed based on the 2018 World Cancer Research Fund/American Institute for Cancer Research Cancer Prevention Recommendations and included nine dietary and lifestyle components: fruit and vegetable intake, total fiber intake, red meat intake, alcohol consumption, physical activity, body mass index (BMI), waist circumference (WC), smoking status, and sugar-sweetened drink intake. HLIS was categorized into tertiles (low, intermediate, high) according to its distribution among non-cases. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated using multivariable Cox proportional hazards models with age as the time metric. The parsimonious model was adjusted for family history of breast cancer, educational attainment, type of menopause, parity, daily energy intake, PRS, and the top 10 principal components (PCs), while the extended model was additionally adjusted for history of diabetes, age at menopause, age of menarche, and hormone use. Analyses were further stratified by PRS (above vs. below the median), and interaction was tested using a likelihood ratio test. Results: Among the 22,725 postmenopausal women (non-Hispanic White [22.2%], African American [12.9%], Native Hawaiian [8.1%], Japanese American [37.5%], and Latino [19.2%]), 1,174 developed breast cancer during an average follow-up of 12.2 years. Higher HLIS was associated with a significantly lower risk of breast cancer. Compared to women in the low HLIS tertile, breast cancer risk was 12% lower (HR = 0.88; 95% CI: 0.77-1.00; P = 0.05) in the intermediate tertile and 34% lower (HR = 0.66; 95% CI: 0.56-0.77; P < 0.001) in the high tertile, with similar results in the extended model. This inverse association appeared to be driven by BMI (obese vs. normal: HR=1.36, 95% CI: 1.13-1.64, P=0.001) and WC (≥88 cm vs. <80 cm: HR=1.43, 95% CI = 1.19-1.72, P < 0.001). PRS was significantly associated with breast cancer risk (per-SD HR = 1.37; 95% CI: 1.29-1.45; P < 0.001). The inverse association between HLIS and breast cancer risk was similar across PRS strata, with no significant interaction (P-int = 0.48). Conclusion: In this multiethnic cohort of postmenopausal women, adherence to a healthy lifestyle was associated with a reduced risk of breast cancer, regardless of low or high PRS, supporting the importance of lifestyle modification as a key preventive strategy for breast cancer across diverse populations. Citation Format: Chenya Zhao, Gertraud Maskarinec, David Conti, Christopher A. Haiman, Loïc Le Marchand, Lynne R. Wilkens, Fei Chen, Eunjung Lee. The impact of genetic and lifestyle factors on the risk of invasive postmenopausal breast cancer in the multiethnic cohort study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 5035.

Associations of dietary scores (per 1-SD increase) with prostate cancer among men with PSA screening information in the MEC (N = 62,151), with and without adjustment for PSA screening history.

Supplementary Table S5. Results from multinomial logistic regression models to evaluate pathway-based polygenic risk score interactions with red meat and processed meat intake by topological tumor location.

Associations of dietary scores (per 1-SD increase) with prostate cancer risk among Native Hawaiians in the MEC.

Supplementary Table S6. Associations between red meat/processed meat intake with colorectal cancer risk stratified by quartiles of pathway-based polygenic risk scores.

Abstract Background: The HSD3B1 gene encodes an androgen synthesis enzyme critical for prostate cancer (PCa) progression. The CC genotype of a common missense variant in HSD3B1 (rs1047303) is considered adrenal-permissive, resulting in increased androgen synthesis, resistance to androgen deprivation therapy (ADT), and accelerated PCa progression. However, epidemiologic evidence linking this variant to prostate cancer-specific mortality (PCSM), particularly in diverse patient populations, remains limited. Methods: We investigated the recessive effect of the C allele (CC vs. AA or AC) on PCSM among incident PCa cases in the Multiethnic Cohort (MEC). Cause-specific Cox proportional hazards regression models were used to estimate the hazard ratios (HRs) and 95% CI confidence intervals (CIs), with age since PCa diagnosis as the time scale. Covariates included tumor stage (localized, regional, or distant), Gleason grade (high: ≥8 or low: <8), first-degree family history of PCa, the first ten principal components, and initial course of treatment (chemotherapy, radiation therapy, hormone therapy, or surgery). Analyses were conducted overall and in cases stratified by tumor stage, grade, and metastatic status. A sensitivity analysis was conducted, restricted to men who underwent hormone therapy as part of the initial treatment. Statistical significance was defined as a two-sided p-value <0.05. Results: Among the 3,216 incident PCa cases (34.7% Japanese Americans, 20.4% Latinos, 19.5% African Americans, 19.1% Whites, and 6.4% Native Hawaiians), 295 (9.2%) died due to PCa over a median follow-up of 5.9 years, and 115 (3.6%) carried the CC genotype. Compared to the AA/AC genotype, the CC genotype was associated with a suggestive 57% increased risk of PCSM (95% CI: 0.80-3.09, P=0.19). The association was stronger among patients with metastatic (HR=5.24, 95% CI: 1.01-8.48, P=0.01), advanced (HR=2.33, 95% CI: 0.67-8.08, P=0.18), or high-grade (HR=2.15, 95% CI: 0.80-5.79, P=0.13) PCa. In the hormone therapy subgroup, the CC genotype was significantly associated with increased PCSM among patients with metastatic (HR=6.79, 95% CI: 2.03-17.2, P<0.001) and advanced (HR=3.29, 95% CI: 1.10-11.7, P=0.03) disease. Conclusion: In this multiethnic population of PCa cases, the adrenal-permissive CC genotype of HSD3B1 rs1047303 variant was associated with a markedly increased risk of PCSM, particularly among patients with aggressive disease and those treated with hormone therapy. These findings suggest that HSD3B1 genotyping may have clinical utility in identifying patients at higher risk of poor outcomes and guiding personalized treatment strategies. Further studies with larger sample sizes are needed to confirm these associations and explore their implications for clinical decision-making. Citation Format: Wei Xiong, Xin Sheng, Peggy Wan, Lynne R. Wilkens, Loïc Le Marchand, David V. Conti, Christopher A. Haiman, Fei Chen. Adrenal-permissive HSD3B1 genotype and prostate cancer-specific mortality among patients: Insights from the multiethnic cohort Study [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3594.

Distribution of dietary scores by PSA screening history in the MEC.

There is a growing interest in evaluating the intersection of genetic and environmental factors, particularly social determinants of health (SDoH). As both the distributions and associations of genetic and SDoH-related risk vary across populations, a thorough understanding of the interplay of these factors (genetics and SDoH across populations) is necessary for the appropriate design and interpretation of studies examining their combined impact on health outcomes. In this review, we review population descriptors, including self-reported social constructs and genetically defined constructs, highlighting the different concepts they may capture and when it may be appropriate to use them. We discuss the challenges of applying polygenic risk scores (PRSs) to populations distinct in their genetic architecture or social context from the cohort in which they were developed. We provide an overview of conceptual SDoH frameworks and measures at the individual and area levels, discussing how these measures are defined, assessed, utilized, and interpreted in health research. For evaluating SDoH and PRS jointly, we outline analytic considerations, including calculating main-effect estimates, conducting gene-environment interaction studies, testing for mediation, and incorporating these factors into clinical prediction algorithms. When examining across populations, we highlight opportunities and challenges of data harmonization across existing cohorts and biobanks and ethical considerations necessary before embarking on or reporting work in this field. In all cases, we highlight the criticality of basing scientific questions upon well-considered conceptual frameworks arising from prior established relationships between risk factors and disease.

Supplementary Table S4. Pathway annotation, mapped genes, and putative effect for single nucleotide polymorphisms that were overrepresented in PANTHER pathways.

Pairwise correlation coefficients between each dietary score examined in analyses.