Causal relationship between diet and common female diseases: a comprehensive Mendelian randomization study

Utilizing a two-sample MR framework, we assessed the causal effects of diet on nine common female-related diseases. The study design and the underlying MR hypotheses are illustrated in Fig. 1 . Fig. 1 Flowchart of MR analysis in this study. SNP, single nucleotide polymorphism. Flowchart of MR analysis in this study. SNP, single nucleotide polymorphism. This two-sample MR study relied on publicly available GWAS pooled statistics, and ethical approval was obtained for the original study. The data sources used in this study are summarized in Table  1 . We extracted the major SNPs associated with diet and female-related diseases and used them as genetic IVs for the two-sample MR (Supplementary Tables S1-32). To mitigate the heterogeneity issues associated with the use of public databases, we restricted our analysis to individuals of European ancestry. Uniform and rigorous quality control procedures were applied across all datasets, including harmonization of effect alleles, exclusion of palindromic SNPs, clustering based on a consistent linkage disequilibrium reference panel, as well as testing and correction for directional horizontal pleiotropy using the MR-Egger method. The Q statistic from the inverse-variance weighted (IVW) method, along with its corresponding p-value, is reported to quantify the extent of heterogeneity. Additionally, leave-one-out sensitivity analysis was conducted by iteratively removing each SNP to assess whether the overall findings were driven by any single SNP, potentially originating from heterogeneous data sources. The GWAS data on diet- and female-associated diseases are stored in the Integrated Epidemiology Unit (IEU) Open GWAS project ( https://gwas.mrcieu.ac.uk/ ). Table 1 Details of the data sources used in this study Exposure or Outcome GWAS ID Consortium Sample size Number of SNPs Diet  Shellfish intake ukb-b-143 MRC-IEU 64,939 9,851,867  Fresh fruit intake ukb-b-3881 MRC-IEU 446,462 9,851,867  Coffee intake ukb-b-5237 MRC-IEU 428,860 9,851,867  Pork intake ukb-b-5640 MRC-IEU 460,162 9,851,867  Alcohol intake frequency ukb-b-5779 MRC-IEU 462,346 9,851,867  Pancake intake ukb-b-6500 MRC-IEU 64,949 9,851,867  Yogurt intake ukb-b-7753 MRC-IEU 64,949 9,851,867  Omelette intake ukb-b-11272 MRC-IEU 64,949 9,851,867  Herbal tea intake ukb-b-13344 MRC-IEU 64,949 9,851,867  Lamb/mutton intake ukb-b-14179 MRC-IEU 460,006 9,851,867  Lobster/crab intake ukb-b-14746 MRC-IEU 64,938 9,851,867  Soya dessert intake ukb-b-998 MRC-IEU 64,947 9,851,867  Dark chocolate intake ukb-b-16139 MRC-IEU 64,945 9,851,867  Dried fruit intake ukb-b-16576 MRC-IEU 421,764 9,851,867  Non-oily fish intake ukb-b-17627 MRC-IEU 460,880 9,851,867  Cheese intake ukb-b-1489 MRC-IEU 451,486 9,851,867  Salad / raw vegetable intake ukb-b-1996 MRC-IEU 435,435 9,851,867  Oily fish intake ukb-b-2209 MRC-IEU 460,443 9,851,867  Whole-wheat cereal intake ukb-b-2375 MRC-IEU 64,949 9,851,867  White pasta intake ukb-b-2512 MRC-IEU 64,949 9,851,867  Beef intake ukb-b-2862 MRC-IEU 461,053 9,851,867  Milk intake ukb-b-2966 MRC-IEU 64,943 9,851,867 Gynaecological Disease  Breast cancer (Oncoarray) ieu-a-1129 BCAC 106,776 10,680,257  Cervical cancer ieu-b-4876 NA 199,086 8,506,261  Endometrial cancer ebi-a-GCST006464 NA 121,885 9,470,555  Ovarian cancer ieu-a-1120 OCAC 66,450 0  Endometriosis ebi-a-GCST90018839 NA 231,771 24,089,752  Ovarian cyst ebi-a-GCST90018889 NA 218,469 24,100,412  Polycystic ovary syndrome (adjusted for age) ebi-a-GCST90044902 NA 141,355 22,981,890  Primary ovarian failure finn-b-E4_OVARFAIL NA — 16,379,677  Uterine fibroids ebi-a-GCST90018934 NA 258,718 24,129,853 Note: MRC-IEU, Medical Research Council Integrative Epidemiology Unit;BCAC,Breast Cancer Association Consortium;OCAC,Ovarian Cancer Association Consortium Details of the data sources used in this study Note: MRC-IEU, Medical Research Council Integrative Epidemiology Unit;BCAC,Breast Cancer Association Consortium;OCAC,Ovarian Cancer Association Consortium Instrumental variables were selected according to the following standardized procedure: (1) Genome-wide Significance: SNPs significantly associated with each exposure at the conventional genome-wide threshold ( P < 5e-08) were initially identified. (2) Linkage Disequilibrium Clumping: To ensure independence among IVs, we performed LD-based clumping using a reference panel from the 1000 Genomes Project European samples. SNPs with an R² < 0.01 within a 5,000 kb window were clustered, retaining only the variant with the smallest P -value. (3) Palindromic SNPs: Palindromic SNPs with intermediate allele frequencies were excluded to prevent strand ambiguity during harmonization. (4) Instrument Strength: We computed the F-statistic for each SNP. All IVs displayed F-statistics substantially greater than 10, indicating a low risk of weak instrument bias. In addition, we estimated the proportion of variance explained (R²) by each individual SNP using summary-level data (beta, standard error, and effect allele frequency) from the source GWAS. For dietary traits with limited genome-wide significant SNPs (e.g., alcohol intake frequency, beef intake), we applied a relaxed threshold ( P < 5e-06) when fewer than five independent SNPs were available at the conventional threshold. This approach is consistent with prior MR studies on dietary factors [ 16 ]. Even under this relaxed criterion, all LD clumping and quality control procedures were rigorously maintained. To enhance confidence in findings derived from these instruments, we prioritized results that were consistent across multiple MR methods (IVW, weighted median, and MR-Egger) and confirmed the absence of significant horizontal pleiotropy using the MR-Egger intercept test [ 17 ]. This study was guided by the Guidelines for Enhancing Reporting of Observational Studies in Epidemiology Using Mendelian Randomization (STROBE-MR) [ 18 ]. Three core assumptions must be satisfied to conduct MR analyses to derive valid causal estimates of the association between the exposure of interest and disease outcome: [ 1 ] genetic variation is strongly associated with the exposure of interest; [ 2 ] genetic variation is not associated with any known or potential confounders; and [ 3 ] genetic variation is not associated with the outcome in any way other than through the exposure [ 19 ]. We utilized a two-sample MR analysis framework, where GWAS summary statistics from different exposure and outcome studies were used to estimate causal effects. This approach enhances the statistical power and precision of the MR analysis. We conducted MR analyses using the TwoSampleMR R package (version 0.5.7) [ 20 ]. Our primary analysis utilized the inverse variance weighted (IVW) method to estimate the association between dietary habits and the risk of female-specific diseases. For palindromic variation, we inferred positive stranded alleles using allele frequencies, setting the behavioral stringency to action = 2 in the harmonise_data function. To evaluate pleiotropy, we performed MR-Egger regression [ 21 ]. In cases of horizontal pleiotropy, we opted to eliminate confounders and re-run the MR analysis. We further employed the MR-PRESSO method to identify and remove potential outliers [ 21 ]. We then assumed that the remaining variation contributing to the genetic instrumentation was valid [ 22 ], to provide unbiased estimates. To account for multiple testing across the primary exposure–outcome pairs, we applied the Benjamini–Hochberg procedure to control the false discovery rate (FDR). Associations with an FDR-corrected P -value < 0.05 were considered statistically significant. Results below the nominal threshold ( P < 0.05) but not surviving FDR correction were interpreted as suggestive and clearly indicated as such. All analyses were performed based on the R program (version 4.2.3) using the “TwoSampleMR” package (version 0.5.7).

Our study identified associations between certain dietary factors and BC, CC, and EC. We found a significant correlation between BC and the consumption of dried fruit and beef (Fig. 2 A and B; Table  2 ; Supplementary Figure S1-2). Increased dried fruit intake was linked to a reduced risk of BC (IVW, OR = 0.54; 95%CI:0.40,0.72; P  = 2.97E-05), while beef consumption was identified as a risk factor for BC (IVW, OR = 1.92; 95%CI:1.13,3.28; P  = 0.02). Fig. 2 Odds ratio plot for BC, CC, EC and exposure factors A BC and beef intake B BC and dried fruit intake C CC and dried fruit intake D EC and alcohol intake frequency; E. EC and non-oily fish intake; F. EC and coffee intake Odds ratio plot for BC, CC, EC and exposure factors A BC and beef intake B BC and dried fruit intake C CC and dried fruit intake D EC and alcohol intake frequency; E. EC and non-oily fish intake; F. EC and coffee intake Table 2 Summary data for this study Expourse Outcome Method P -Value OR FDR- P -Value Dried fruit intake Breast cancer (Oncoarray) Inverse variance weighted 2.97E-05 0.539088417 0.029403 Beef intake Inverse variance weighted 0.015311876 1.928587542 0.524314211 Dried fruit intake Cervical cancer Inverse variance weighted 0.014167426 0.993442478 0.524314211 Coffee intake Endometrial cancer Inverse variance weighted 0.015769195 1.743581864 0.524314211 Alcohol intake frequency Inverse variance weighted 0.00195941 1.272733333 0.223807716 Non-oily fish intake Inverse variance weighted 0.047074351 1.613443877 0.795311157 Pork intake Ovarian cancer Inverse variance weighted 0.001689273 0.529797321 0.223807716 Alcohol intake frequency Inverse variance weighted 0.001763584 1.221607868 0.223807716 Dried fruit intake Inverse variance weighted 0.009423089 0.602833654 0.490992532 Non-oily fish intake Inverse variance weighted 0.004115937 1.778222834 0.314331321 Beef intake Inverse variance weighted 0.026045389 0.490966355 0.622944419 Shellfish intake Endometriosis Inverse variance weighted 0.031060443 3.450860362 0.640621637 Coffee intake Inverse variance weighted 0.003403682 0.449426076 0.30633138 Cheese intake Inverse variance weighted 0.005543179 0.608653475 0.365849814 Coffee intake Ovarian cyst Inverse variance weighted 0.000183449 0.555128337 0.06053817 Omelette intake Inverse variance weighted 0.046838386 1.557488891 0.795311157 Lobster/crab intake Inverse variance weighted 0.010481394 0.236061756 0.500353213 Dark chocolate intake Inverse variance weighted 0.024532529 0.643276333 0.622944419 Dried fruit intake Inverse variance weighted 0.00554021 0.589431762 0.365849814 Whole-wheat cereal intake Inverse variance weighted 0.018527764 1.730126215 0.524314211 Herbal tea intake Primary ovarian failure Inverse variance weighted 0.018335635 1.050232555 0.524314211 Soya dessert intake Inverse variance weighted 0.029574129 0.000338641 0.622944419 Dark chocolate intake Inverse variance weighted 0.014755677 26.98419463 0.524314211 Whole-wheat cereal intake Inverse variance weighted 0.000646047 209.5947726 0.159896633 Fresh fruit intake Polycystic ovary syndrome Inverse variance weighted 0.002125384 4.086475402 0.223807716 Yogurt intake Inverse variance weighted 0.002260684 3.604880384 0.223807716 Herbal tea intake Inverse variance weighted 0.031823963 1.016859094 0.642973946 Coffee intake Uterine fibroids Inverse variance weighted 7.71E-05 1.812728755 0.0381645 Dried fruit intake Inverse variance weighted 0.002031839 1.519974515 0.223807716 Non-oily fish intake Inverse variance weighted 0.029287696 1.31019623 0.622944419 Summary data for this study In contrast, CC was solely associated with dried fruit intake in our study (Fig. 2 C; Table  2 ; Supplementary Figure S3 ), which was also identified as a protective factor reducing the risk of CC (IVW, OR = 0.99; 95%CI:0.9882,0.9987; P  = 0.01) We also discovered dietary links to EC, another prevalent gynecologic cancer. The consumption of coffee (IVW, OR = 1.74; 95%CI:1.11,2.74; P  = 0.02), frequency of alcohol intake (IVW, OR = 1.27; 95%CI:1.09,1.48; P  = 0.002), and non-oily fish intake (IVW, OR = 1.61; 95%CI:1.006,2.59; P  = 0.047) were all identified as risk factors for EC occurrence (Fig. 2 D-F; Table  2 ; Supplementary Figure S4-6). Our MR analysis identified evidence supporting a potential causal relationship between dietary factors and OC (Fig. 3 ; Table  2 ; Supplementary Figure S7-11). Frequent alcohol consumption (IVW, OR = 1.22; 95%CI:1.08,1.38; P  = 0.002) and non-oily fish intake (IVW, OR = 1.78; 95%CI:1.2,2.63; P  = 0.004) have been identified as risk factors for OC. Conversely, consumption of pork (IVW, OR = 0.53; 95%CI:0.36,0.79; P  = 0.002), dried fruit (IVW, OR = 0.60; 95%CI:0.41,0.88; P  = 0.009), and beef (IVW, OR = 0.49; 95%CI:0.26,0.92; P  = 0.03) serve as protective factors against OC. Fig. 3 Odds ratio plot for OC and exposure factors Odds ratio plot for OC and exposure factors The causal association between diet and OCS is even more pronounced (Fig. 4 ; Table  2 ; Supplementary Figure S12-17). Consumption of omelettes (IVW, OR = 1.56; 95%CI:1.01,2.41; P  = 0.047) and whole-wheat cereals (IVW, OR = 1.73; 95%CI:1.10, 2.73; P  = 0.02) are risk factors for OCS. On the other hand, coffee (IVW, OR = 0.56; 95%CI:0.41,0.76; P  = 0.0002), lobster/crab (IVW, OR = 0.24; 95%CI:0.08,0.71; P  = 0.01), dark chocolate (IVW, OR = 0.64; 95%CI:0.44,0.94; P  = 0.02), and dried fruit (IVW, OR = 0.59; 95%CI:0.41,0.86; P  = 0.006) intake decrease the likelihood of OCS, thus acting as protective factors. Fig. 4 Odds ratio plot for OCS and exposure factors Odds ratio plot for OCS and exposure factors A causal relationship has been established between ES and shellfish intake (IVW, OR = 3.45; 95%CI:1.12,10.64; P  = 0.03), coffee consumption (IVW, OR = 0.445; 95%CI:0.26,0.77; P  = 0.003), and cheese intake (IVW, OR = 0.61; 95%CI:0.43,0.86; P  = 0.006) (Fig. 5 A-C; Table  2 ; Supplementary Figure S18-20). Shellfish intake increases the risk of ES, while coffee and cheese serve as protective factors. Uterine fibroids (UF), the most common benign tumor in women, are primarily associated with coffee (IVW, OR = 1.81; 95%CI:1.35,2.43; P  = 7.71E-05), dried fruit (IVW, OR = 1.52; 95%CI:1.16,1.98; P  = 0.002), and non-oily fish (IVW, OR = 1.31; 95%CI:1.03,1.67; P  = 0.03) intake, all of which are risk factors for UF occurrence (Fig. 5 D-F; Table  2 ; Supplementary Fig. 21–23). Fig. 5 Odds ratio plot for ES, UF and exposure factors A ES and coffee intake B ES and cheese intake C ES and shellfish intake D UF and coffee intake E UF and non-oily fish intake F UF and dried fruit intake Odds ratio plot for ES, UF and exposure factors A ES and coffee intake B ES and cheese intake C ES and shellfish intake D UF and coffee intake E UF and non-oily fish intake F UF and dried fruit intake Polycystic ovary syndrome (PCOS), another common ovarian-related disease, was primarily associated with the intake of fresh fruit (IVW, OR = 4.09; 95%CI:1.66,10.03; P  = 0.002), yogurt (IVW, OR = 3.60; 95%CI:1.58, 8.21; P  = 0.002), and herbal tea (IVW, OR = 1.02; 95%CI:1.001,1.03; P  = 0.03). All these factors were identified as risk factors for the development of PCOS (Fig. 6 A; Table  2 ; Supplementary Fig. 24–26). Fig. 6 Odds ratio plot for PCOS, POF and exposure factors A PCOS and diet B POF and soya dessert intake C POF and dark chocolate intake D POF and whole-wheat cereal intake E POF and herbal tea intake Odds ratio plot for PCOS, POF and exposure factors A PCOS and diet B POF and soya dessert intake C POF and dark chocolate intake D POF and whole-wheat cereal intake E POF and herbal tea intake The association of POF was observed with the intake of soya dessert (IVW, OR = 0.0003; 95%CI:2.53E-07,0.45; P  = 0.03), herbal tea (IVW, OR = 1.05; 95%CI:1.01,1.09; P  = 0.02), dark chocolate (IVW, OR = 26.98; 95%CI:1.91,381.49; P  = 0.01), and whole-wheat cereal (IVW, OR = 209.59; 95%CI:9.72,4519.33; P  = 0.0006). Among these, soya dessert intake was identified as a protective factor while the others were found to be risk factors (Fig. 6 B-E; Table  2 ; Supplementary Fig. 27–30).

The current clinical management of female-specific diseases relies heavily on surgical interventions. For instance, gynecological malignancies are typically treated with surgery followed by postoperative radiotherapy. Although early-stage gynecological cancers often show high long-term survival rates, these aggressive procedures considerably compromise women’s quality of life and overall health. Even benign gynecological conditions can profoundly affect daily functioning and well-being. In this context, our Mendelian randomization (MR) study focuses on the early prevention of such diseases. Recognizing diet as a readily modifiable lifestyle factor, we aimed to evaluate whether dietary modifications could mitigate the risk of female-specific diseases, thereby contributing to improved quality of life. In summary, our study applied MR methodology to systematically assess the causal relationships between dietary factors and nine common female-specific diseases. We found that favorable dietary changes could substantially reduce the risk of several of these conditions. Therefore, we recommend that healthcare providers integrate evidence-based dietary guidance into routine women’s health care, with the goal of reducing disease risk and enhancing long-term quality of life and well-being. A key limitation of this study is its restriction to European-ancestry populations, which constrains the generalizability of our findings. Genetic architecture, lifestyle patterns, and social determinants of health vary substantially across racial, ethnic, and geographic groups. These differences may influence both dietary behaviors and disease mechanisms, meaning that causal estimates derived from European populations may not be directly applicable to other groups. Furthermore, we acknowledge that sample overlap between the dietary and gynecological disease GWAS sources (primarily from the UK Biobank and FinnGen) represents another methodological consideration. In this study, we mitigated this concern by ensuring all instrumental variables exhibited high F-statistics (well above 10) and by relying on the consistency of results across several MR methods (IVW, weighted median, and MR-Egger) with different underlying assumptions. Nevertheless, the potential influence of sample overlap on the confidence intervals of our estimates should be noted. Additionally, our study employed univariable MR, which assumes that the genetic instruments affect the outcome only through the exposure of interest. However, many dietary factors are correlated. Therefore, we cannot fully rule out that the observed associations for one dietary factor might be partly confounded by other related aspects of diet. Future genome-wide association studies that include more diverse cohorts—and capture dietary phenotypes and genetic variants relevant to female diseases across populations—will be essential. With broader genetic resources, future MR analyses could employ stronger instrumental variables and larger sample sizes and would be well-positioned to apply multivariable MR to disentangle the direct and independent causal effects of correlated dietary exposures. Such advances will help produce more reliable and broadly generalizable causal inferences.

This MR study used large-scale GWAS data to investigate potential causal links between diet and common female diseases. This method minimizes confounding and reverse causation, offering more robust causal inference than conventional observational studies. After FDR correction, only two associations remained significant: dried fruit intake as a protective factor for BC, and coffee intake as a risk factor for UF. These represent the most reliable findings. All other exposure-outcome associations showed only nominal significance ( P   0.05) and should be considered exploratory, suggesting hypotheses for future research rather than confirming causality. These nominal associations—involving foods like dried fruit, non-oily fish, and various meats—could be true signals our study was underpowered to confirm, or false positives. For some dietary traits (e.g., omelette, beef), genetic instruments were selected with a relaxed threshold ( P  < 5e-6) due to limited SNP availability, increasing the risk of weak instrument bias and false positives. These results require caution. Other limitations include potential residual pleiotropy, sample overlap, and the restriction to European-ancestry populations, limiting generalizability. Future work should validate the two significant associations in diverse populations and clarify their mechanisms. For the nominal findings, larger GWAS with better genetic instruments are needed. In summary, this study provides robust evidence that nut intake may protect against BC and coffee may increase UF risk. The numerous nominal associations highlight the complexity of diet-disease relationships and the need for further research with improved methods. Dietary patterns have been found to significantly influence the risk of cancer. Numerous observational [ 23 , 24 ] have established a correlation between dietary patterns and the risk of BC. MR studies also have identified dried fruit as a protective factor against BC development [ 25 ]. This finding aligns with our MR analysis results, suggesting that individuals who regularly consume dried fruit have a lower risk of BC compared to those who do not. A review by Laudisio, D. et al. [ 26 ] proposed that fruits and vegetables, olive oil, fish, and red wine, all rich in antioxidant properties, may contribute to a reduced risk of BC. However, our MR analysis did not establish a causal link between these daily dietary patterns and BC incidence. Interestingly, we found a positive causal relationship between beef consumption and BC, indicating that individuals who consume more beef have a higher risk of developing BC. This is supported by a cohort study by Parada, H. et al. [ 27 ], which reported an association between red meat consumption prior to diagnosis and increased BC-specific mortality. A recent review [ 28 ] also suggested that increased red meat consumption, including beef, may elevate the risk of BC. Research has demonstrated that grilling and frying generate elevated levels of heterocyclic amines (HAAs), and that meat smoked or grilled over open flames or on heated surfaces contains polycyclic aromatic hydrocarbons (PAHs). HAAs, which exhibit genotoxic properties, are absorbed in the human gastrointestinal tract. PAHs have been shown to promote DNA adduct formation and disrupt apoptotic processes. Although several studies have explored the mechanisms through which meat consumption increases the risk of colon cancer, fundamental research in this area remains limited for BC [ 28 ]. Therefore, further research is necessary to elucidate the mechanisms involved in this association. The prevailing perspective posits that CC is precipitated by certain types of Human Papillomavirus (HPV) infections. Reactive oxygen species (ROS) may contribute to the mechanisms underlying cervical carcinogenesis. Prior observational studies have proposed that the ingestion of dietary antioxidants could neutralize detrimental ROS, thereby potentially slowing or preventing persistent HPV infections and reducing the likelihood of CC [ 29 ]. In our analysis, genetically predicted dried fruit intake was associated with a lower risk of CC (IVW, OR = 0.99; 95%CI:0.9882,0.9987; P = 0.01). However, it is crucial to note that the effect estimate is exceedingly close to the null value, and the confidence intervals only narrowly exclude one. This suggests that while the association is statistically significant, the magnitude of the effect is likely negligible from a biological or public health perspective. Furthermore, such a fragile association may be susceptible to minor biases or unmeasured confounding. Therefore, we interpret this finding with considerable caution and emphasize that it should be considered as hypothesis-generating, requiring robust replication in future studies before any substantive conclusions can be drawn. This finding aligns with a previous MR analysis that identified dried fruit intake as a protective factor against CC [ 25 ]. Unlike fresh fruits, dried fruits are preferred due to their longer shelf life and reduced susceptibility to nutrient loss. They are rich in a diverse array of bioactive components and phytochemicals that can modulate the onset and progression of numerous chronic diseases by influencing metabolic pathways and cellular responses, thereby promoting health and longevity [ 30 ]. However, research on the correlation between dried fruit intake and cancer remains incomplete, necessitating further exploration of the potential mechanisms through which dried fruits may impact cancer. Endometrial Cancer (EC) is common in older postmenopausal women and is characterized by abnormal postmenopausal uterine bleeding. Current research on the causal relationship between coffee intake and EC presents conflicting results [ 31 , 32 ]. Yang, T. O. et al. found no causal link, and at best, a weak association between the risk of EC and coffee consumption in a cohort study and meta-analysis involving British women [ 31 ]. Conversely, Lafranconi, A. et al. demonstrated through a meta-analysis that coffee intake reduces the risk of postmenopausal EC [ 32 ]. Interestingly, our MR analysis revealed a positive association between coffee consumption and EC, suggesting an increased risk of EC. The role of alcohol consumption in the development of EC has been well-documented in several studies [ 33 , 34 ]. However, the relationship between alcohol consumption and EC appears to be complex, with some studies suggesting that alcohol may actually reduce the risk of EC in certain populations, such as overweight women [ 34 , 35 ]. Our MR analysis also found a positive association between alcohol consumption and EC, suggesting that alcohol may increase the risk of EC. Therefore, further epidemiological studies are needed to better understand this relationship. Omega-3 fatty acids, commonly found in fish, are often touted for their health benefits, including a potential protective effect against cancer. However, the evidence supporting this claim is mixed. Some animal and in vitro studies have found a link between omega-3 fatty acid consumption and reduced cancer risk [ 36 , 37 ], but epidemiological studies have produced conflicting results [ 38 – 40 ]. Our MR analysis found a positive correlation between non-oily fish intake and EC, suggesting that increased intake of non-oily fish may be a risk factor for EC. However, we did not find a causal link between oily fish intake and EC. Therefore, more research is needed to confirm these findings. Ovarian cancer (OC) remains a major health concern for women worldwide, characterized by frequently late diagnosis and poor prognosis. Identifying modifiable risk factors and developing effective prevention and early detection strategies are therefore critical. Current evidence on the relationship between dietary factors and OC remains limited and often inconsistent [ 41 ]. For instance, while one case-control study reported a significant inverse association between poultry consumption and OC risk [ 42 ], a finding partially aligned with our MR results suggesting pork and beef as protective—the same study identified fish as a protective factor, contradicting our MR estimates. These discrepancies underscore the need for more refined epidemiological and mechanistic studies to clarify the causal role of specific dietary components in OC. Notably, our MR analysis identified dried fruit consumption as a potential protective factor for OC. Experimental evidence demonstrates that methyl lucidone, a bioactive constituent of dried fruits, exerts dose-dependent cytotoxic effects on OC cells. Its anti-proliferative mechanism involves the induction of G2/M phase cell cycle arrest and the promotion of apoptosis through concurrent activation of the intrinsic apoptotic pathway and suppression of PI3K/Akt survival signaling [ 43 ]. Furthermore, a novel fig-derived compound has been shown to synergistically enhance trastuzumab’s efficacy by boosting phagocytic cell activity via Fcγ receptor binding [ 44 ]. These experimental findings provide mechanistic validation and biological plausibility for prior MR studies, corroborating a protective role of dried fruit intake. The convergence of epidemiological and molecular evidence underscores the translational potential of investigating bioactive dietary metabolites for OC prevention and therapy. In contrast, the relationship between fish intake and OC risk remains contentious. While meta-analyses have suggested that high intakes of eicosapentaenoic and docosahexaenoic acids—abundant in oily fish—are associated with reduced OC risk [ 36 ], and that higher fish consumption is linked to lower mortality among OC patients [ 45 ], our MR analysis indicated a potential risk-increasing effect of non-oily fish intake. This divergence suggests that the type of fish, preparation methods, or environmental contaminants may modulate the association, underscoring the limitations of broad dietary categorization in nutritional epidemiology. Regarding alcohol consumption, although earlier meta-analyses reported no significant association with OC [ 46 ], a recent case-control study [ 47 ] suggested that higher intake may elevate OC risk—a finding consistent with our MR results. This shift in evidence may reflect improved study designs, more accurate exposure assessment, or the inclusion of genetically predisposed subpopulations, pointing to alcohol as a potential modifiable risk factor worthy of further investigation. Endometriosis, a hormone-dependent disease, is a leading cause of dysmenorrhea and infertility in women. Presently, dietary modification has also been suggested as a potential therapeutic approach for ES. Previous observational studies and meta-analyses have indicated that caffeine intake does not appear to be linked with an increased risk of ES [ 48 , 49 ]. However, our MR analysis revealed a negative association between coffee consumption and ES, suggesting a protective role of coffee against the disease. A meta-analysis [ 50 ] showed that a higher intake of high-fat dairy products and cheese might be associated with a reduced risk of ES. This finding is supported by a case-control study on Iranian women [ 51 ], which reported a significant correlation between the consumption of cheese and a lower risk of ES. These observational study results align with our MR findings, which also suggested a protective role of cheese intake against ES. Shellfish consumption has been reported to reduce inflammation and prevent various chronic non-communicable diseases. However, the relationship between shellfish intake and health outcomes remains controversial [ 36 , 45 ]. A prospective cohort study [ 52 ] found no association between shellfish intake and the risk of developing ES. In contrast, our MR analysis indicated a positive association between shellfish consumption and ES, suggesting shellfish as a potential risk factor for the disease. Ovarian cysts, predominantly benign, are categorized into physiological and pathological types. Our MR analysis revealed a strong correlation between OCS and dietary habits. The correlation between OCS and diet has been relatively underexplored, with only one prior case-control study indicating a higher fat intake among women with OCS [ 53 ]. Our MR analysis revealed a negative association between OCS and the consumption of coffee, lobster/crab, dark chocolate, and dried fruit, suggesting these foods may serve as protective factors. Conversely, consumption of omelettes and whole-wheat cereal was positively correlated with OCS, indicating potential risk factors. This underscores the need for further observational studies to elucidate the causal relationship between diet and OCS. Premature ovarian failure is characterized by amenorrhea before the age of 40 due to ovarian failure. The link between diet and ovarian-related diseases is well-established, yet no observational studies have investigated the causal relationship between diet and POF. Our MR analysis showed a positive association between soya dessert intake and POF, while herbal tea, dark chocolate, and whole-wheat cereal intake were negatively associated with POF. This highlights the need for more comprehensive in vivo and in vitro studies to understand the underlying mechanisms. Polycystic ovary syndrome, a multifaceted hormonal disorder prevalent in women of reproductive age, has been associated with obesity, insulin resistance, and hyperandrogenemia, and extrinsic factors such as stress and diet have also been implicated [ 54 ]. Dietary therapy is a key treatment approach for PCOS, reflecting the strong causal relationship between PCOS and insulin resistance. High-fat foods increase the risk of insulin resistance-related diseases, while foods rich in unsaturated fatty acids mitigate this risk. In our MR study, we observed a positive correlation between yogurt intake and PCOS, suggesting yogurt as a potential risk factor for PCOS. This finding, however, contradicts a previous study by Li, T. et al. [ 55 ], which demonstrated that inulin-enriched yogurt mitigated reproductive dysfunction in a PCOS mouse model. Consequently, further in vitro and in vivo studies are warranted to elucidate the precise mechanism of yogurt’s impact on PCOS. In a recent case-control study [ 56 ] involving 1854 participants, fruit consumption was identified as an independent risk factor for PCOS. This finding aligns with our MR study, although the specific mechanisms remain unclear. Existing research indicates that green tea, oolong tea, sage tea, and marjoram herb possess protective properties in PCOS patients [ 57 , 58 ]. For instance, green tea enhances ovulation, reduces cyst formation in PCOS patients, and lowers plasma corticosterone levels and uterine contractility in patients with dysmenorrhea [ 59 ]. Similarly, catechins in oolong tea suppress uterine inflammation and matrix degradation, thereby mitigating ovarian dysfunction and insulin resistance in PCOS mice [ 58 ]. However, our MR study identified “Herbal tea intake” as a risk factor for PCOS, a finding that contradicts the protective role observed for other teas. This discrepancy necessitates further clinical trials to validate the causal relationship between herbal tea consumption and PCOS. Uterine fibroids, are benign growths that originate in the uterus, primarily composed of smooth muscle cells and fibrous connective tissue. Research suggests that estrogen level imbalances may contribute to their formation. A study by Eugster, H. P. et al. [ 60 ] proposed potential interactions between caffeine and estrogen metabolism, warranting further investigation. An earlier cross-sectional study reported that caffeine intake elevated early follicular estradiol levels [ 61 ]. Despite this biological evidence, caffeine intake has not been established as a risk factor for UF [ 62 ]. Our MR study identified a strong correlation between coffee consumption and UF, suggesting coffee intake as a potential risk factor for UF development. To date, only a limited number of studies have examined the relationship between diet and UF. An Italian case-control study indicated that high consumption of beef, red meat, and ham was associated with an increased risk of UF, while fish intake was associated with a reduced risk [ 62 ]. Conversely, a retrospective study [ 63 ] found that fish intake of polychlorinated biphenyls was positively associated with the risk of UF, contradicting the findings of Wise et al. [ 64 ], who found no association between the risk of UF and fish and seafood intake. In our MR study, non-oily fish intake was positively associated with the risk of UF development, suggesting it as a potential risk factor. The current evidence on the relationship between fish consumption and UF is conflicting, necessitating further studies to elucidate their causal relationship. Unlike the existing risk factors associated with UF, our MR study also revealed a significant association between dried fruit consumption and the risk of developing UF. However, it is important to note that no observational studies have established a causal relationship between the two. Therefore, further research is necessary to explore the potential causal link between dried fruit consumption and UF.

The global burden of cancer, in terms of both incidence and mortality, continues to rise markedly, establishing cancer as a leading cause of death worldwide [ 1 , 2 ]. This increasing trend in incidence, mortality, and overall disease burden among females is largely driven by breast, cervical, ovarian, and endometrial cancers, which rank among the most common causes of cancer-related deaths in women. Concurrently, although non-malignant gynecological conditions—such as uterine fibroids (UF), polycystic ovary syndrome (PCOS), premature ovarian failure (POF), and ovarian cysts (OCS)—may not directly lead to mortality, they substantially impair women’s quality of life [ 3 , 4 ]. Consequently, a key focus of current research is the development of strategies to mitigate the impact of these female-specific diseases. Diet and nutrition represent a fundamental pillar of population health, with relevance to women’s health. Nutritional deficiencies can adversely affect health status, and conversely, health conditions may influence nutritional requirements. The interaction between genetic predisposition and nutritional factors plays a critical role in health maintenance and disease prevention. Nutrients can modulate gene expression and influence susceptibility to various diseases, including cancer, through multiple biological mechanisms [ 5 – 8 ]. Gynecological diseases encompass a spectrum of disorders affecting the female reproductive system, including benign and malignant tumors, infectious processes, and endocrine abnormalities. Benign conditions such as uterine fibroids (UF) and endometriosis (ES) are highly prevalent and exert a considerable negative impact on quality of life. In contrast, gynecological malignancies have emerged in recent years as one of the leading causes of female mortality. A substantial body of observational research has explored the relationship between dietary patterns and gynecological diseases [ 8 , 9 ]. For instance, although the association between breast cancer (BC) and specific foods or nutrients has been extensively studied, the findings remain inconsistent. A meta-analysis by D. Aune et al. reported only a marginal inverse correlation between fruit and vegetable consumption and BC risk [ 10 ]. Another study by Teresa T. Fung et al. suggested that higher intakes of berries and peaches may lower the risk of estrogen receptor-negative BC among postmenopausal women [ 11 ]. In contrast, the Italian cohort of the European Prospective Investigation into Cancer and Nutrition found no significant association between fruit consumption and BC incidence [ 12 ]. We posit that these discrepancies may arise from synergistic or antagonistic interactions among dietary components. Furthermore, conventional observational studies are often limited by residual confounding—due to unmeasured or imprecisely measured variables—reverse causation, and other biases that complicate causal interpretation. In recent years, Mendelian randomization (MR) has gained prominence as a method for strengthening causal inference. MR employs single-nucleotide polymorphisms (SNPs) as instrumental variables (IVs) that are independent of common confounding factors. By leveraging the random assignment of genetic variants at conception, MR mimics a randomized controlled trial, thereby reducing susceptibility to confounding. Moreover, because genetic variants are generally fixed before disease onset, MR minimizes the risk of reverse causation, offering a more robust framework for evaluating potential causal relationships. Although several MR studies have been conducted on the association between diet and gynecological diseases, the existing literature is predominantly concentrated on malignant tumors [ 13 , 14 ]. For instance, causal effects of micronutrients on BC and EC have been evaluated in separate MR investigations [ 13 , 14 ], while the relationship between ovarian cancer and diet has been examined only for a narrow range of dietary factors, such as coffee and alcohol [ 15 ]. Nevertheless, the application of MR to non-neoplastic gynecological pathologies has been limited. In the present study, we therefore employed a comprehensive two-sample MR framework to investigate the potential causal relationships between varied dietary patterns and a wide array of female-specific diseases. Our study aims to use genetic variants as instrumental variables to assess the potential causal relationship between dietary factors and female-related cancers, offering new insights for disease prevention.

Supplementary Material 1. Supplementary Figure S1. Causal relationship between “Beef intake” and breast cancer. A) Scatter plot. The slope of each line represents the causal relationship of each method. B) Forest plot. The significance of red lines are MR results of MR-Egger test and IVW method. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. The blue line represents the IVW estimate, and the dark blue line represents the MR-Egger estimate.Supplementary Figure S2. Causal relationship between “Dried fruit intake” and breast cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S3. Causal relationship between “dried fruit intake” and cervical cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity.Supplementary Figure S4. Causal relationship between “Non-oily fish intake” and endometrial cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S5. Causal relationship between “Alcohol intake frequency” and Endometrial cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S6. Causal relationship between “Coffee intake” and Endometrial cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S7. Causal relationship between “dried fruit intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S8. Causal relationship between “Alcohol intake frequency” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S9. Causal relationship between “Beef intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S10. Causal relationship between “Non-oily fish intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S11. Causal relationship between “Pork intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S12. Causal relationship between “Whole-wheat cereal intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S13. Causal relationship between “Dried fruit intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S14. Causal relationship between “Dark chocolate intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S15. Causal relationship between “Lobster/crab intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S16. Causal relationship between “Omelette intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S17. Causal relationship between “Coffee intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S18. Causal relationship between “Cheese intake” and Endometriosis. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S19. Causal relationship between “Coffee intake” and Endometriosis. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S20. Causal relationship between “Shellfish intake intake” and Endometriosis. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S21. Causal relationship between “dried fruit intake” and Uterine fibroid. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S22. Causal relationship between “Non-oily fish intake” and Uterine fibroid. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S23. Causal relationship between “Coffee intake” and Uterine fibroid. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S24. Causal relationship between “Herbal tea intake” and Polycystic ovary syndrome. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S25. Causal relationship between “Yogurt intake” and Polycystic ovary syndrome. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S26. Causal relationship between “Fresh fruit intake” and Polycystic ovary syndrome. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S27. Causal relationship between “Herbal tea intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S28. Causal relationship between “Soya dessert intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S29. Causal relationship between “Dark chocolate intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S30. Causal relationship between “Whole-wheat cereal intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Material 1. Supplementary Figure S1. Causal relationship between “Beef intake” and breast cancer. A) Scatter plot. The slope of each line represents the causal relationship of each method. B) Forest plot. The significance of red lines are MR results of MR-Egger test and IVW method. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. The blue line represents the IVW estimate, and the dark blue line represents the MR-Egger estimate.Supplementary Figure S2. Causal relationship between “Dried fruit intake” and breast cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S3. Causal relationship between “dried fruit intake” and cervical cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity.Supplementary Figure S4. Causal relationship between “Non-oily fish intake” and endometrial cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S5. Causal relationship between “Alcohol intake frequency” and Endometrial cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S6. Causal relationship between “Coffee intake” and Endometrial cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S7. Causal relationship between “dried fruit intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S8. Causal relationship between “Alcohol intake frequency” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S9. Causal relationship between “Beef intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S10. Causal relationship between “Non-oily fish intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S11. Causal relationship between “Pork intake” and Ovarian cancer. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S12. Causal relationship between “Whole-wheat cereal intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S13. Causal relationship between “Dried fruit intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S14. Causal relationship between “Dark chocolate intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S15. Causal relationship between “Lobster/crab intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S16. Causal relationship between “Omelette intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S17. Causal relationship between “Coffee intake” and Ovarian cysts. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S18. Causal relationship between “Cheese intake” and Endometriosis. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S19. Causal relationship between “Coffee intake” and Endometriosis. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S20. Causal relationship between “Shellfish intake intake” and Endometriosis. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S21. Causal relationship between “dried fruit intake” and Uterine fibroid. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S22. Causal relationship between “Non-oily fish intake” and Uterine fibroid. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S23. Causal relationship between “Coffee intake” and Uterine fibroid. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S24. Causal relationship between “Herbal tea intake” and Polycystic ovary syndrome. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S25. Causal relationship between “Yogurt intake” and Polycystic ovary syndrome. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S26. Causal relationship between “Fresh fruit intake” and Polycystic ovary syndrome. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S27. Causal relationship between “Herbal tea intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S28. Causal relationship between “Soya dessert intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S29. Causal relationship between “Dark chocolate intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Figure S30. Causal relationship between “Whole-wheat cereal intake” and Premature ovarian failure. A) Scatter plot. B) Forest plot. C) Leave-one-out plot. D) Funnel plot to assess heterogeneity. Supplementary Material 2. Supplementary Table S1. IVs for Shellfish intake (P < 5e-08).Supplementary Table S2. IVs for Fresh fruit intake (P < 5e-08).Supplementary Table S3. IVs for Coffee intake (P < 5e-08).Supplementary Table S4. IVs for Pork intake (P < 5e-08).Supplementary Table S5. IVs for Alcohol intake frequency. (P < 5e-06).Supplementary Table S6. IVs for Pancake intake (P < 5e-08).Supplementary Table S7. IVs for Yogurt intake (P < 5e-08).Supplementary Table S8. IVs for Omelette intake (P < 5e-06).Supplementary Table S9. IVs for Herbal tea intake (P < 5e-08).Supplementary Table S10. IVs for Lamb/mutton intake (P < 5e-08). Supplementary Table S11. IVs for Lobster/crab intake (P < 5e-08).Supplementary Table S12. IVs for Soya dessert intake (P < 5e-08).Supplementary Table S13. IVs for Dark chocolate intake (P < 5e-08).Supplementary Table S14. IVs for Dried fruit intake (P < 5e-08).Supplementary Table S15. IVs for Non-oily fish intake (P < 5e-08).Supplementary Table S16. IVs for Cheese intake (P < 5e-06).Supplementary Table S17. IVs for Salad / raw vegetable intake (P < 5e-08).Supplementary Table S18. IVs for Oily fish intake (P < 5e-08).Supplementary Table S19. IVs for Whole-wheat cereal intake (P < 5e-06).Supplementary Table S20. IVs for White pasta intake (P < 5e-08).Supplementary Table S21. IVs for Beef intake (P < 5e-06).Supplementary Table S22. IVs for Milk intake (P < 5e-08).Supplementary Table S23. Potential confounders of all SNPs in the PhenoScanner database.Supplementary Table S24. MR Results of diet on Breast cancer.Supplementary Table S25. MR Results of diet on Cervical cancer.Supplementary Table S26. MR Results of diet on Endometrial cancer.Supplementary Table S27. MR Results of diet on Ovarian cancer.Supplementary Table S28. MR Results of diet on Endometriosis.Supplementary Table S29. MR Results of physical activity on Ovarian cyst.Supplementary Table S30. MR Results of diet on Polycystic ovary syndrome.Supplementary Table S31. MR Results of diet on Primary ovarian failure.Supplementary Table S32. MR Results of diet on Uterine fibroids. Supplementary Material 2. Supplementary Table S1. IVs for Shellfish intake (P < 5e-08).Supplementary Table S2. IVs for Fresh fruit intake (P < 5e-08).Supplementary Table S3. IVs for Coffee intake (P < 5e-08).Supplementary Table S4. IVs for Pork intake (P < 5e-08).Supplementary Table S5. IVs for Alcohol intake frequency. (P < 5e-06).Supplementary Table S6. IVs for Pancake intake (P < 5e-08).Supplementary Table S7. IVs for Yogurt intake (P < 5e-08).Supplementary Table S8. IVs for Omelette intake (P < 5e-06).Supplementary Table S9. IVs for Herbal tea intake (P < 5e-08).Supplementary Table S10. IVs for Lamb/mutton intake (P < 5e-08). Supplementary Table S11. IVs for Lobster/crab intake (P < 5e-08).Supplementary Table S12. IVs for Soya dessert intake (P < 5e-08).Supplementary Table S13. IVs for Dark chocolate intake (P < 5e-08).Supplementary Table S14. IVs for Dried fruit intake (P < 5e-08).Supplementary Table S15. IVs for Non-oily fish intake (P < 5e-08).Supplementary Table S16. IVs for Cheese intake (P < 5e-06).Supplementary Table S17. IVs for Salad / raw vegetable intake (P < 5e-08).Supplementary Table S18. IVs for Oily fish intake (P < 5e-08).Supplementary Table S19. IVs for Whole-wheat cereal intake (P < 5e-06).Supplementary Table S20. IVs for White pasta intake (P < 5e-08).Supplementary Table S21. IVs for Beef intake (P < 5e-06).Supplementary Table S22. IVs for Milk intake (P < 5e-08).Supplementary Table S23. Potential confounders of all SNPs in the PhenoScanner database.Supplementary Table S24. MR Results of diet on Breast cancer.Supplementary Table S25. MR Results of diet on Cervical cancer.Supplementary Table S26. MR Results of diet on Endometrial cancer.Supplementary Table S27. MR Results of diet on Ovarian cancer.Supplementary Table S28. MR Results of diet on Endometriosis.Supplementary Table S29. MR Results of physical activity on Ovarian cyst.Supplementary Table S30. MR Results of diet on Polycystic ovary syndrome.Supplementary Table S31. MR Results of diet on Primary ovarian failure.Supplementary Table S32. MR Results of diet on Uterine fibroids. Supplementary Material 3. Supplementary Material 3. Supplementary Material 4. Supplementary Material 4.