Abstract
18
STUDY QUESTION 19
Can a large-scale genome-wide association study (GWAS) meta-analysis identify the genomic 20
risk loci and associated candidate genes for female genital tract (FGT) polyps, provide insights 21
into the mechanism underlying their development , and inform potential overlap with other 22
traits, including endometrial cancer? 23
SUMMARY ANSWER 24
GWAS meta-analysis of FGT polyps highlighted the potentially shared mechanisms between 25
polyp development and cancerous processes. 26
WHAT IS KNOWN ALREADY 27
Small-scale candidate gene studies have focused on biological processes such as estrogen 28
stimulation and inflammation to clarify the biology behind FGT polyps. However, the exact 29
mechanism for the development of polyps is still elusive. At the same time, a genome -wide 30
approach, which has become the gold standard in complex disease genetics, has never been 31
used to uncover the genetics of the FGT polyps. 32
STUDY DESIGN, SIZE, DURATION 33
We performed a genome wide association study (GWAS) meta -analysis including a total of 34
25,100 women with FGT polyps (International Classification of Disease, ICD -10 diagnosis 35
code N84) and 207,193 female controls (without N84 code) of European ancestry from the 36
FinnGen study (11,092 cases and 94,394 controls) and the Estonian Biobank (EstBB, 14,008 37
cases and 112,799 controls). 38
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
3
PARTICIPANTS/MATERIALS, SETTING, METHODS 39
A meta-analysis and functional annotation of GWAS signals were performed to identify and 40
prioritise genes in associated loci. To determine associations with other phenotypes, we 41
performed a look-up of associated variants across multiple traits and health conditions, a genetic 42
correlation analysis, and a phenome -wide association study (PheWAS) with ICD10 diagnosis 43
codes. 44
MAIN RESULTS AND THE ROLE OF CHANCE 45
Our GWAS meta-analysis revealed ten significant (P < 5 x 10-8) genomic risk loci. Two signals, 46
rs2277339 (P = 7.6 x 10-10) and rs1265005 (P = 1.1 x 10-9) (in linkage disequilibrium (LD) with 47
rs805698 r2 = 0.75), are exonic missense variants in PRIM1, and COL17A1 genes, respectively. 48
Based on the literature, these genes may play a role in cellular proliferation. Several of the 49
identified genomic loci had previously been linked to endometrial cancer and/or uterine 50
fibroids. Thus, highlighting the potentially shared mechanisms underlying tissue overgrowth 51
and cancerous processes, which may be relevant to the development of polyps. Genetic 52
correlation analysis revealed a negative correlation between sex hormone -binding globulin 53
(SHBG) and the risk of FGT polyps (rg = -0,21, se = 0.04, P = 2.9 x 10-6), and on the phenotypic 54
level (PheWAS), the strongest associations were observed with endometriosis, leiomyoma of 55
the uterus and excessive, frequent and irregular menstruation. 56
LARGE SCALE DATA 57
The complete GWAS summary statistics will be made available after publication through the 58
GWAS Catalogue (https://www.ebi.ac.uk/gwas/). 59
LIMITATIONS, REASONS FOR CAUTION 60
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
4
In this study, we focused broadly on polyps of FGT and did not differentiate between the polyp 61
subtypes. The prevalence of FGT polyps led us to assume that most women included in the 62
study had endometrial polyps. Further study on the expression profile of FGT polyps could 63
complement the GWAS study to substantiate the functional importance of the identified 64
variants. 65
WIDER IMPLICATIONS OF THE FINDINGS 66
The study findings have the potential to significantly enhance our understanding of the genetic 67
mechanisms involved, paving the way for future functional follow -up, which in turn could 68
improve the diagnosis, risk assessment, and targeted treatment options, since surgery is the only 69
line of treatment available for diagnosed polyps. 70
TRIAL REGISTRATION NUMBER: 71
Not applicable 72
Key words: Genome-wide association study, Female genital tract polyps, Endometrial polyps, 73
Benign disorders, cell proliferation, PRIM1, COL17A1. 74
Introduction
75
Polyps of the female genital tract (FGT) are generally benign tissue overgrowths found 76
in both reproductive-aged and postmenopausal women. While the prevalence of polyps can be 77
as high as 50%, most women are asymptomatic. Hence, polyps are usually detected incidentally 78
during routine ultrasound examinations, diagnostic hysteroscopy for other gynaecological 79
disorders, or infertility treatment among women of reproductive age (Hinckley and Milki, 2004; 80
Fatemi et al., 2010; Karayalcin et al., 2010; Bettocchi et al., 2011). However, the occurrence 81
of polyps is age-dependent, with a higher prevalence among postmenopausal women compared 82
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
5
to premenopausal women (Dreisler et al., 2009). Despite the asymptomatic and benign nature 83
of FGT polyps, for some women, they can negatively impact the daily quality of life by causing 84
abnormal vaginal bleeding and infertility. 85
Endometrial polyps (EPs) are the most common type of FGT polyps, with a prevalence 86
ranging from 7.8% to 50% (Dreisler et al., 2009; de Azevedo et al., 2016; Tanos et al., 2017). 87
In contrast, endocervical polyps occur only in 2% to 5% of cases , whereas vaginal polyps are 88
rarer (Tanos et al. , 2017) ; therefore, limited evidence is available about their biology and 89
clinical implications. Some known risk factors for the development of EPs are advanced age, 90
hypertension, diabetes, obesity, hyper oestrogenism, administration of hormone replacement 91
therapy (HRT), and tamoxifen (Vitale et al., 2021; Vieira et al., 2022). Nevertheless, the exact 92
pathogenesis of FGT polyps remains unclear. Histologically, EPs are characterised by large, 93
thickened blood vessels with fibrous stroma and irregularly shaped glandular spaces (Tanos et 94
al., 2017). Moreover, EPs can be associated with concomitant intrauterine pathologies like 95
endometrial hyperplasia, adenomyosis, endometriosis and chronic endometritis (Annan et al., 96
2012; Raz et al., 2021). Despite the benign nature of EPs, around 3.5% of cases progress into 97
carcinoma (Lee et al., 2010; Sasaki et al., 2018; Uglietti et al., 2019). At the same time, the 98
extent of biological mechanisms and pathways shared between EPs and carcinoma, as well as 99
unique molecular characteristics of each condition remain unclear. 100
Very little is known about the heritability and genetic background of FGT polyps. 101
Genetic factors, including chromosomal translocations in 6p21-22, 12q13-15, or 7q22 regions, 102
may contribute to polypoid morphology (Dal Cin et al., 1995; Nijkang et al., 2019; Vieira et 103
al., 2022), and genetic disorders like Lynch or Cowden syndrome have been reported to be 104
accountable for the development of EPs (Vieira et al., 2022). However, a recent study by Sahoo 105
et al. did not confirm the presence of chromosomal rearrangements in endometrial polyps 106
(Sahoo et al., 2022). 107
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
6
Thus, the confirmed pathophysiology for the development of FGT polyps is still 108
unknown. However, it can be assumed that the development and mechanism of polyps are 109
multifactorial and can also depend on genetic predisposition. Thus far, t he proposed genetic 110
mechanisms are primarily based on small candidate gene studies and are inconclusive (Altaner 111
et al., 2006; Pal et al., 2008; Banas et al., 2018; Doria et al., 2018; Takeda et al., 2019). Studies 112
in other complex diseases have shown that genome-wide association studies (GWAS) can 113
provide valuable insight into disease biology (Claussnitzer et al., 2016), but as far as we are 114
aware, no large-scale GWAS have been published for FGT polyps. Therefore, we performed a 115
GWAS meta-analysis to identify the genetic variants associated with FGT polyps, followed by 116
numerous post-GWAS analyses to understand the shared and unique genetic underpinnings of 117
FGT polyps and endometrial cancer. Ultimately, this knowledge can highlight potential clinical 118
avenues for appropriate diagnosis and management of FGT polyps. 119
Materials and methods
120
Ethical approval 121
All Estonian Biobank (EstBB) participants have signed a broad informed consent form, 122
and analyses were carried out under ethical approvals 1.1 -12/624 and 1.1 -12/2733 from the 123
Estonian Committee on Bioethics and Human Research (Estonian Ministry of Social Affairs) 124
and data release application 6-7/GI/630 from the EstBB. For the FinnGen study, we used only 125
publicly available GWAS summary statistics without individual-level data and thus, a separate 126
ethics approval was not needed. 127
Study cohorts 128
Our analyses included a total of 25,100 women with polyps of the FGT (International 129
Classification of Disease, ICD -10 diagnosis code N84) and 207,193 female controls (without 130
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
7
the N84 code) of European ancestry from two studies: summary level statistics from the 131
FinnGen R7 data release (11,092 cases and 94,394 controls) and individual-level data from the 132
EstBB (14,008 cases and 112,799 controls). In FinnGen, cases were defined using the ICD -10 133
code N84 or corresponding ICD-9 (6210, 6227, 6237, 6246) and ICD-8 (62520) codes. Similar 134
to EstBB, controls were defined as women who did not have the abovementioned disease codes. 135
To increase study power, we did not distinguish between different types of FGT polyp 136
location in our phenotype definition. However, according to the FinnGen Risteys browser 137
(https://r7.risteys.finngen.fi/phenocode/N14_POLYPFEMGEN), 20% of FGT polyp cases 138
were also cases for phenotype “Uterine polyps”, while in the EstBB data, 70.5 % of N84 cases 139
had a diagnosis for uterine polyps (as defined by the presence of the ICD-10 code N84.0). This 140
most likely reflects differences in the source of phenotype information - FinnGen phenotype 141
definitions mostly use hospital records and thus involve more severe cases, while the EstBB 142
also uses primary care records. 143
Cohort-level analyses 144
The EstBB is a population -based biobank including more than 200,000 individuals 145
(approximately 135,000 of them women) representing 20% of the Estonian adult population. 146
Information on ICD codes is obtained via regular linking with the National Health Insurance 147
Fund and other relevant databases. Individuals with ICD-10 code N84 (mean age at recruitment 148
49.2 years, standard deviation 12.1) were categorised as cases having been diagnosed with 149
polyps of the genital tract, and all female biobank participants without the diagnosis were 150
considered as controls (mean age at recruitment 44.3 years, sd 16.6). The genotyping procedure 151
for EstBB has been described previously (Koel et al., 2023; Laisk et al., 2021; Pujol-Gualdo et 152
al., 2022). Briefly, Illumina GSAv1.0, GSAv2.0, and GSAv2.0_EST arrays were used for the 153
genotyping of biobank participants at the Core Genotyping Lab of the Institute of Genomics, 154
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
8
University of Tartu. Individuals with a call -rate less than 95% and sex mismatch between 155
phenotypic and chromosomal data were excluded from the analysis. Before imputation, 156
genotyped variants were filtered by applying criteria of call rate < 95%, Hardy -Weinberg 157
equilibrium P < 10 −4 (for autosomal variants only), and minor allele frequency < 1%. Eagle 158
v2.3 software was used for pre-phasing, and Beagle was used for imputation. The population -159
specific imputation reference of 2297 whole genome sequencing samples was used. Association 160
analysis was performed using REGENIE v2.2.4 with year of birth and 10 principal components 161
as covariates in step I, and variants with a minor allele count < 5 were excluded by default. In 162
downstream association analysis, poorly imputed variants with an INFO score < 0.4 were 163
excluded. 164
For FinnGen, GWAS summary statistics from the R7 data release were used, and 165
therefore, individual-level data was not available. The summary statistics were obtained from 166
https://www.finngen.fi/en/access_results, whereas the FinnGen cohort and the genotyping/data 167
analysis details have been previously described in Kurki et al. (Kurki et al. , 2023) . To 168
summarise, age, 10 principal components, and genotyping batch were used in REGENIE v2.0.2 169
analysis as covariates, and for FinnGen summary statistics, variants with a minor allele count 170
> 5 and imputation INFO score > 0.6 were included. 171
GWAS meta-analysis 172
A meta-analysis using fixed -effects inverse variance weighting with genomic control 173
was performed with the GWAMA v2.1 tool. The genome -wide significance level was set to 174
P < 5 × 10-8. The genomic inflation factors (lambda) for the individual study summary statistics 175
were 1.046 (EstBB) and 1.045 (FinnGen). Variants present in both cohorts (n = 12,363,169) 176
were included in downstream analyses. 177
Annotation of GWAS signals 178
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
9
To annotate the GWAS signals and prioritise potential biologically relevant genes at 179
associated loci, we adopted the following approach. 180
First, we used the Functional Mapping and Annotation of Genome -Wide Association 181
Studies (FUMA GWAS) platform v1.5.2 to identify genetic association loci. FUMA is an online 182
platform designed for the annotation , prioritisation, and interpretation of GWAS results that 183
uses data from multiple databases to annotate GWAS signals (Watanabe et al., 2017). In the 184
first step of this analysis, independent significant variants, lead signals and genomic risk loci 185
were defined. Independent significant variants (IndSigSNPs) were defined as variants that were 186
genome-wide significant (P < 5 x 10-8) and had a pairwise LD r2 < 0.6, according to the 1000G 187
EUR reference panel. From this subset, lead variants were derived. Finally, risk loci were 188
defined from independent significant SNPs by merging LD blocks if they are less apart than r2 189
< 0.6. Thus, a genomic risk locus can contain several lead SNPs and/or independent significant 190
SNPs, depending on the size of the locus and the LD structure. Thereafter, potential significant 191
candidate SNPs (GWAS meta-analysis P > 0.05) were determined to be in LD with any of the 192
IndSigSNPs (r2 ≥ 0.6) within a 1Mb window and had a MAF of ≥ 1%. These candidate SNPs 193
were subjected to further annotation using multiple databases such as Annotate Variation 194
(ANNOVAR) (Wang et al. , 2010), RegulomeDB scores (ranging from 1 to 7, where lower 195
score indicates greater evidence for having regulatory function) (Boyle et al. , 2012) , and 196
Combined Annotation -Dependent Depletion ( CADD) (a continuous score showing how 197
deleterious the SNP is to protein structure/function; scores >12.37 indicate potential 198
pathogenicity) (Kircher et al. , 2014) , 15 chromatin states from the Roadmap Epigenomics 199
Project (ENCODE Project Consortium, 2012; Roadmap Epigenomics Consortium et al., 2015), 200
expression quantitative trait loci (eQTL) data (genotype -tissue expression (GTEx) v6 and v7) 201
(GTEx Consortium, 2013) and 3D chromatin interactions from HI -C experiments of 21 202
tissues/cell types (Schmitt et al. , 2016) . This process provides information on the location, 203
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
10
functional impact, and potential regulatory effects of detected SNPs. For eQTL annotations, we 204
focused on tissues similar to uterine/vaginal tissue (GTEx Consortium, 2020). 205
Using the lead signal identified by FUMA, we performed a look-up in the Open Targets 206
genetics database (Ghoussaini et al., 2021). This database combines several layers of evidence 207
across a wide range of cell types and tissues to generate an aggregate ‘variant to gene’ (V2G) 208
score. The V2G score provides identification and correlation of likely causal variants and genes 209
to prioritise the potential functional genes associated with the identified variants. V2G score 210
aggregates several parameters like distance from the canonical transcript start site, eQTLs and 211
protein QTLs (pQTLs) datasets, datasets for chromatin interactions and conformation, 212
molecular phenotypes, and in silico functional predictions (using the Variant Effect Predictor 213
or VEP score). 214
Briefly, we prioritised genes in the identified loci by selecting three genes having the 215
highest V2G score. Then we additionally identified those loci where the GWAS signal includes 216
a missense variant using annotation data from FUMA. To provide additional support for 217
prioritisation and further explore regulatory effects, we looked at eQTL associations according 218
to FUMA annotations, and if none were reported, we looked at potential chromatin interactions 219
as these may also indicate regulatory effects. To gain insight into endometrial-specific eQTLs, 220
we queried the candidate SNPs defined by FUMA in the endometrial eQTL database (Fung et 221
al., 2018). 222
Gene-based testing 223
Analysis by Multi-marker Analysis of GenoMic Annotation (MAGMA) v1.6 (de Leeuw 224
et al., 2015), with the default settings in FUMA (Watanabe et al., 2017), was used to perform 225
gene-based association analysis to complement the single variant analyses. Gene-based analysis 226
enables to detect the joint effect of multiple genetic variants and can thus increase the power to 227
detect associations. Briefly, variants located in the gene body were assigned to protein -coding 228
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
11
genes (n = 18,895; Ensembl build 85), and the SNP P -values were merged into a gene test 229
statistic using the SNP -wise mean model (de Leeuw et al., 2015). The level of genome -wide 230
significance was set at 0.05/18,895 = 2.6 x 10-6, taking into account the number of tested genes. 231
Genetic associations with other traits: 232
During the FUMA functional mapping, candidate SNPs were linked with the GWAS 233
catalogue (https://www.ebi.ac.uk/gwas/, GWAS Catalogue e0_r2022-11-29, FUMA v1.5.1) to 234
explore the association of genetic variants with previously published GWAS of different 235
phenotypic traits. 236
Genetic correlation analysis 237
To estimate the genetic correlations between our FGT polyps meta -analysis and 1,335 238
other traits, we utilised the LD Score Regression method implemented in the Complex Traits 239
Genetics Virtual Lab (CTG-VL) (https://genoma.io/) and 3 additional endometrial cancer traits 240
available in the GWAS catalog ue (accession codes GCST006464, GCST006465 and 241
GCST006466) (Bulik-Sullivan et al. , 2015) . Statistical significance was determined by 242
applying a multiple testing correction (FDR < 5%) using the p.adjust function in R v3.6.3. 243
Associations with other phenotypic traits 244
To determine the associations between ICD -10 diagnosis main codes and the N84 245
diagnosis of polyps in the FGT, we conducted an analysis using individual level data from the 246
EstBB. Logistic regression was used to test the associations between N84 and other ICD -10 247
codes while controlling for age at recruitment and 10 genetic principal components to account 248
for population stratification and avoid false associations due to ancestry/regional differences 249
between cases and controls. Age was included as a covariate to account for incomplete 250
electronic diagnosis data for older participants in the biobank. Our analysis was limited to the 251
diagnosis main codes along with all the subcodes to increase the power of the data. Statistically 252
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
12
significant associations were determined by applying Bonferroni correction (2000 tested ICD 253
main codes, corrected P threshold of 2.5 x 10 -5). Odds ratios (ORs) were calculated by the 254
logistic regression method and represented as adjusted ORs. The resulted associations were 255
filtered to remove the diagnoses related to exogenous factors (such as injuries, poisoning, 256
accidents, assaults, etc. in the S, T, U, V, W, X, and Y subchapters) and the PheWAS library 257
(https://github.com/PheWAS/PheWAS) was used to visualise the results. All analyses were 258
performed using R v4.1.3 259
In the FinnGen data, we did not have access to the individual level data, but using the 260
Risteys portal ( https://r7.risteys.finngen.fi/phenocode/N14_POLYPFEMGEN), we explored 261
the results of the survival analysis evaluating associations between FGT polyps and other 262
selected phenotypes. A detailed description of the survival analysis can be found in the Risteys 263
documentation (https://r7.risteys.finngen.fi/documentation), but briefly, this type of analysis 264
tests the association between an exposure endpoint and an outcome endpoint. For example, in 265
the context of FGT polyps, what is the association between a diagnosis of FGT polyp (exposure 266
endpoint) and endometrial carcinoma (outcome endpoint). 267
Results
268
Summary of GWAS 269
GWAS of a total of 25,100 women with polyps of the FGT and 207,193 female controls 270
revealed ten significant (P < 5 x 10 -8) genomic risk loci and 23 independent signals (Fig. 1, 271
Table 1). The strongest signal, rs1702136 (P = 2.17 x 10 -19), was observed on chromosome 3, 272
which is an intronic variant of the EEFSEC gene. The majority of the significant signals 273
(91.3%) were either intronic or intergenic variants, whereas two signals, rs2277339 (P = 7.57 x 274
10-10) and rs1265005 (P = 1.09 x 10-9, in LD with rs805698, r2 = 0.75) on chromosomes 12 and 275
10, respectively, were exonic missense variants. 276
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
13
277
Table 1 . Summary statistics of significant genomic risk loci with lead SNP and 278
independent significant SNP. 279
chr:pos:A1:A2 rsID of
Lead SNP
(effect
allele)
P-value OR (95%
CI)
IndSigSNP
s
Effect
Allele
frequency
(Est/Fin)
Heter
ogene
ity p -
value
3:128118711:A:
G
rs1702136
(G)
2.17 x 10 -
19
0.90 (0.88-
0.92)
rs1702136
rs3732402
rs4857866
rs7650365
rs2999051
rs13095166
rs74924715
rs1735527
0.75/0.75 0.99
8:116869477:C:
T
rs800578
(T)
4.91 x 10 -
11
1.08 (1.05-
1.10)
rs800578
rs2736213
0.24/0.26 0.90
4:95731394:A:G rs2865375
(A)
4.36 x 10 -
10
0.93 (0.92-
0.95)
rs2865375
rs10033997
0.54/0.52 0.45
12:57146069:G:
T
rs2277339(
T)
7.57 x 10 -
10
1.10 (1.06-
1.13)
rs2277339 0.87/0.86 0.37
10:105585753:C
:T
rs19309775
3 (C)
7.59 x 10 -
10
0.83 (0.78-
0.88)
rs19309775
3
rs17116149
rs7911816
rs11360949
6
rs1265005
0.97/0.95 0.92
1:61592380:A:G rs12751005
(A)
1.09 x 10-9 1.06 (1.04-
1.09)
rs12751005 0.68/0.67 0.90
13:41869725:A:
G
rs77478686
(G)
8.21 x 10-9 1.18 (1.12-
1.26)
rs77478686 0.97/0.95 0.67
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
14
5:142162633:A:
G
rs7728894
(G)
1.84 x 10-8 0.93 (0.91-
0.96)
rs7728894 0.73/0.75 0.60
5:1285974:A:C rs7705526
(C)
2.33 x 10-8 0.94 (0.92-
0.96)
rs7705526 0.68/0.68 0.45
19:8786624:A:G rs2967684
(G)
2.72 x 10-8 0.93 (0.90-
0.95)
rs2967684 0.81/0.76 0.40
*Alleles presented in alphabetical order 280
IndSigSNPs: Independent significant SNP 281
282
283
284
285
286
287
288
289
Figure. 1. Manhattan plot for significant genomic risk loci identified for polyps of female 290
genital tract. On the Manhattan plot, the x -axis represents chromosomes, while the y -axis 291
represents −log10(P-values) for the association of variants identified in polyps of the female 292
genital tract. The horizontal red dashed line represents the genome-wide significance threshold 293
(P < 5 × 10−8). The prioritised genes for each locus are labelled on the top, and genes associated 294
with exonic missense variants are coloured red. 295
Functional annotation of associated variants and gene prioritisation 296
We performed a look -up in the Open Targets Genetics database to evaluate the 297
functional association of identified genomic loci and potentially mapped genes reflected by the 298
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
15
V2G score. We additionally highlighted those loci where the GWAS signal includes a missense 299
variant using annotation data from FUMA (Table 2). Full details of gene prioritisation together 300
with supporting evidence from FUMA eQTL and chromatin interaction mapping, can be found 301
in Supplementary Table 1. 302
Table 2. Summary of gene prioritisation results 303
Locus Lead variant Highest V2G
score
Exonic
variants
Biological evidence
3:128118711:
A:G
rs1702136 EEFSEC
RPN1
DNAJB8
NA EEFSEC potential
endometrial cancer
susceptibility gene
(Kho et al., 2021)
8:116869477:
C:T
rs800578 TRPS1 NA Potential tumour
suppressor candidate
in endometrial cancer
(Liang et al., 2012)
4:95731394:A:
G
rs2865375 BMPR1B
PDLIM5
HPGDS
NA BMPR1B is associated
with female infertility.
BMPR1B-AS1
facilitates endometrial
cancer cell
proliferation (Lai et
al., 2022). BmprIB
mutant mice exhibit a
failure in endometrial
gland formation (Yi et
al., 2001)
12:57146069:
G:T
rs2277339 PRIM1
STAT6
HSD17B6
PRIM1 Cell proliferation,
DNA replication
10:105585753:
C:T
rs193097753 STN1
SH3PXD2A
SLK
COL17A1 Cellular migration,
cellular
differentiation,
extracellular matrix
organisation
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
16
1:61592380:A:
G
rs12751005 NFIA NA NFIA is involved in
cellular growth and
proliferation, tumour
morphology. The gene
is expressed in
endometrial tissue
(Humaidan et al.,
2012)
13:41869725:
A:G
rs77478686 NAA16
MTRF1
KBTBD7
NA Differential expression
of NAA16 is
associated with the
pathogenesis of
endometriosis (She et
al., 2022)
5:142162633:
A:G
rs7728894 ARHGAP26
FGF1
SPRY4
NA ARHGAP26 promotes
ovarian cancer cell
invasion and
migration (Chen et al.,
2019)
5:1285974:A:
C
rs7705526 TERT
CLPTM1L
SLC6A18
NA TERT-CLPTM1L
locus is a known
susceptibility region
for cancerous
processes
19:8786624:A:
G
rs2967684 ACTL9
NFILZ
ADAMTS10
NA ACTL9 locus
associated with
endometriosis
(Rahmioglu et al.,
2023)
304
Based on coding variants in our GWAS signals, we were able to prioritise PRIM1 and COL17A1 305
in loci on chromosomes 12 and 10, respectively. PRIM1 is a DNA primase involved in the 306
initiation of DNA replication by synthesising RNA primers for Okazaki fragments during 307
discontinuous DNA replication (Shiratori et al., 1995). rs2277339 is also a cis-eQTL for PRIM1 308
in endometrial tissue, which further supports PRIM1 as a candidate gene in this locus. 309
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
17
COL17A1, on the other hand, is primarily involved in cellular migration, cellular differentiation, 310
and extracellular matrix organisation (Jones et al., 2020). While these genes have not yet been 311
directly associated with FGT polyps, their biological functions, particularly those related to 312
cellular proliferation, highlight their potential involvement in the development of polyps. 313
Several of the prioritised candidate genes had previously been associated with 314
(endometrial) cancer - EEFSEC, TRPS1, TERT/CLPTM1L, one could be directly linked with 315
endometrial biology ( BMPR1B), and for the remaining, their biological significance in FGT 316
polyps remains unclear. 317
Gene-based associations of female genital tract polyps: 318
To combine the joint effect of multiple genetic variants and increase the power to detect 319
associations, we performed a MAGMA gene -based test implemented in FUMA. Eight genes 320
were identified which passed the recommended threshold for significance (P = 2.6 × 10 -6, 321
Bonferroni correction for association testing of 18,895 protein-coding genes): EEFSEC, TERT, 322
RUVBL1, BMPR1B, TRPS1, COL17A1, BET1L, WBP4 (Supplementary Figure 1, 323
Supplementary Table 2). Majority of these associations mirror the genes prioritised in the single 324
variant analysis, while the BET1L locus was not genome-wide significant in the single variant 325
analysis and is thus novel. Previously, BET1L has been associated with endometrial cancer 326
(Bateman et al., 2017) and uterine fibroids (Cha et al., 2011; Edwards et al., 2013; Liu et al., 327
2018). 328
GWAS catalogue look-up 329
We searched the GWAS catalogue for associations between previously published 330
phenotypic traits and candidate SNPs identified by the FUMA tool to gain additional insight 331
into their potential biological roles (Fig. 2, Supplementary Table 3). Based on the GWAS 332
catalogue look-up, the TERT-CLPTM1L (rs7705526) locus was clearly associated with cancers, 333
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
18
including the development of reproductive cancers such as ovarian cancer and prostate cancer. 334
Potential involvement in cancerous processes was also observed for the EEFSEC locus 335
(rs4857866), which was associated with prostate cancer and also nominally with endometrial 336
cancer (P < 1 x 10 -6). Several identified signals were related to reproductive traits, including 337
menarche (rs2277339-PRIM1), menopause (rs2277339 -PRIM1), gestational age (rs4857866 -338
EEFSEC) and uterine fibroids (rs2277339 -PRIM1, rs193097753 -COL17A1, rs17116149 -339
COL17A1). The rs193097753 signal was additionally associated with the known risk factors 340
for the development of FGT polyps, such as type 2 diabetes (P < 5 x 10-8) and waist-to-hip ratio 341
(P < 2 x 10-11), further substantiating its functional significance. 342
343
344
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
19
345
Figure. 2: GWAS catalogue look -up showing associations between genetic variants 346
associated with female genital tract polyps and other phenotypes. The figure highlights the 347
genome-wide significant (P < 5 x 10 -8) association between FGT polyps genetic risk factors 348
(individual significant SNPs) and previously published phenotypic traits and disorders. 349
Look-up of variants associated with endometrial cancer 350
Since there is some overlap between FGT polyp genetic risk loci and those known to be 351
associated with (endometrial) cancer , we conducted a look -up of variants associated with 352
endometrial cancer (O’Mara et al., 2018) in our GWAS data. Of the 19 SNPs queried, seven 353
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
20
were nominally significant ( P < 0.05) (rs9668337 -SSPN, rs1740828 -SOX4, rs17601876 -354
CYP19A1, rs882380 -SNX11, rs3184504 -SH2B3, rs1129506 -EVI2A, rs11263761 -HNF1B) in 355
the FGT polyp analysis as well (Supplementary Table 4). When querying the 10 lead variants 356
associated with FGT polyp in the endometrial cancer summary statistics, we observed that three 357
were nominally significant ( P < 0.05) in the endometrial cancer and also in the endometrial 358
cancer with endometrioid histology studies (rs1702136 -EEFSEC, rs2865375 -BMPR1B, 359
rs7705526-TERT). 360
Genetic correlation between FGT polyps and other traits 361
We conducted pairwise genetic correlation (rg) analyses to examine the relationship 362
between polyps of the FGT and 1 ,335 different traits obtained from the Complex Traits 363
Genetics Virtual Lab (CTG-VL, https://genoma.io/) and 3 additional endometrial cancer traits 364
available in the GWAS catalogue (accession codes GCST006464, GCST006465 and 365
GCST006466). After running the analysis and correcting for multiple testing, we identified a 366
total of 11 significant (FDR < 0.05) genetic correlations. Selected genetic correlations between 367
polyps and various phenotypic categories such as anthropometrics, cardiovascular diseases, 368
genitourinary traits, mood disorders, sex hormones and surgical procedures are displayed in 369
Fig. 3, full results in Supplementary Table 5 . As expected, a positive genetic correlation was 370
observed with genitourinary traits and hysterectomy, emphasising that polyps are commonly 371
occurring incidental findings in gynaecological disorders. Furthermore, sex hormone -binding 372
globulin (SHBG) was negatively correlated with the risk of FGT polyps. We were not able to 373
detect a significant genetic correlation between FGT polyps and estradiol level or BMI, known 374
risk factors of FGT polyps, in our dataset. FGT polyps showed weak correlation with two of 375
the three endometrial cancer phenotypes tested (endometrial cancer and endometrial cancer - 376
endometrioid histology, both rg 0.25 -0.27, se 0.12 -0.13), but these associations were only 377
nominally significant (P = 0.03) and did not pass multiple testing correction threshold. 378
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
21
379
Figure 3: Results of genetic correlation analysis between endometrial polyps and other 380
relevant traits. 381
The plot displays selected genetic correlations between female genital tract polyps and traits 382
from various phenotypic categories such as anthropometrics, cardiovascular traits, 383
genitourinary traits, mood disorders, sex hormones and surgical procedures that are colour -384
coded. Significant associations after FDR multiple correction are shown in the lower panel, 385
while the upper panel shows selected traits of interest that do not pass the multiple testing 386
correction. The centre dot marks the estimated genetic correlation (rg) value, and error bars 387
indicate 95% confidence limits. The dotted red line indicates no genetic correlation. 388
Associations between polyps of FGT and other diagnosis codes 389
Additionally, to evaluate other phenotypes’ association with FGT polyps, we performed 390
a phenome-wide phenotype association analysis using the EstBB data for disease codes. All 391
239 phenotypic associations with a P < 2.5 x 10 -5 (corresponding to a Bonferroni -corrected 392
threshold of 0.05/2020) are shown in Supplementary Table 6. The most significant association 393
of FGT polyps related to other phenotypes were: i) leiomyoma of uterus (D25 ), which are 394
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
22
uterine fibroids known to often occur together with EPs (Kınay et al., 2016); ii) endometriosis 395
(N80), which has been correlated with a high prevalence of endometrial polyps; iii) excessive, 396
frequent and irregular menstruation (N92) which is one of the classical symptoms of EPs, and; 397
iv) other noninflammatory disorders of uterus except cervix (N85) (Fig. 4). Moreover, women 398
with a diagnosis for FGT polyps had significantly more diagnoses for benign neoplasms in 399
several tissues, such as the uterus (D26), ovary (D27), unspecified female genital organs (D28 400
and D39) and skin (D23) as well as malignant neoplasms of the uterus (C54) and skin (C44). 401
In our analysis, nasal polyps (J33) and K62 (which includes anal/rectal polyps) were also 402
nominally significant, which supports the idea of some systemic tissue overgrowth in epithelial 403
tissues similar to endometrial polyps harbouring glandular and luminal epithelium. 404
For the highlighted associations, we queried the FinnGen Risteys R7 portal to see if we 405
observe similar associations in the Finnish data. While the analysis in the Estonian data does 406
not consider which diagnosis in the tested pair comes first, the FinnGen data survival analysis 407
tests the association in both directions. Uterine leiomyoma, endometriosis, other benign 408
neoplasm of uterus, benign neoplasm of ovary, malignant neoplasm of uterus, and carcinoma 409
in situ of endometrium were significantly associated with FGT polyps in the survival analysis 410
in FinnGen data, (Supplementary Table 7) both if the diagnosis occurred before, or after the 411
FGT polyp diagnosis. 412
The results of phenotypic associations of FGT polyps are in concordance with the observed 413
genetic associations and point to the fact that benign and cancerous processes share some 414
mechanisms (such as cellular proliferation) on a genetic level. 415
416
417
418
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
23
419
420
421
422
423
424
425
426
Figure 4: Association of Polyps of female genital tract (N84) with other phenotypes. 427
Each triangle in the plot corresponds to one ICD10 main code, and different colours represent 428
different diagnosis categories. The direction of the triangle represents the direction of effect and 429
upward-pointing triangles show increased significance of diagnosis code in female genital tract 430
polyps. The red line indicates the Bonferroni corrected threshold for statistical significance. 431
Discussion
432
Gynaecological polyps are a common diagnosis in women, with potential negative 433
implications for women's reproductive health and well -being. We conducted the first large -434
scale GWAS meta -analysis to study the genetic underpinnings of FGT polyps and their 435
association with other phenotypic traits from two European ancestry biobanks. The analysis 436
revealed ten genomic risk loci, of which two (rs2277339 and rs193097753) tagged exonic 437
missense variants suggestive of plausible functional importance in the development of polyps. 438
Furthermore, several of the identified genetic risk loci have previously been associated with 439
(endometrial) cancer and/or uterine fibroids. Genetic correlation analysis additionally showed 440
negative correlation with SHBG levels, but no statistically significant genome-wide correlation 441
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
24
with endometrial cancer. PheWAS analysis showed an increased prevalence of endometriosis, 442
irregular and excessive menstruation, uterine fibroids and neoplasms of the uterus and other 443
tissues in women with FGT polyps. 444
Even though most EPs are benign, there is still a certain level of risk these polyps may 445
progress to malignant transformation (Antunes et al., 2007; Lenci et al., 2014). According to a 446
systematic review and meta -analysis, 2.7% -3.6% of EPs can be considered malignant, with 447
significant heterogeneity in prevalence among pre - and postmenopausal women (Lee et al., 448
2010; Sasaki et al., 2018; Uglietti et al., 2019). While we saw no statistically significant genetic 449
correlation between FGT polyps and endometrial cancer, several of the mapped genetic risk 450
loci/prioritised genes (rs1702136 -EEFSEC, rs800578 -TRPS1, rs2865375-BMPR1B, and 451
rs7705526-TERT-CLPTM1L) have previously been associated with cancer either in GWAS or 452
functional studies (Liang et al., 2012; Wu et al., 2014; Kho et al., 2021; Dos Santos et al., 453
2022), most likely reflecting the fact that both benign and malignant tissue overgrowths may 454
utilise the same cellular mechanisms to some extent. This is supported by the results of the 455
associated diagnoses analysis where women with a diagnosis of FGT polyp also have more 456
diagnoses related to benign and malignant neoplasms of the uterus, ovary, and skin. While the 457
associations with reproductive tract neoplasms may arise due to incidental findings during 458
diagnostic procedures, it cannot be ruled out that some women have a genetic predisposition to 459
tissue overgrowth that may manifest in either benign or malignant neoplasms in certain tissues. 460
Apart from a few small -scale reports (Unler et al. , 2016) , there are no relevant studies 461
evaluating the prevalence of different neoplasms among women with FGT polyps, therefore, 462
our observations need to be confirmed and further evaluated in future studies. 463
We further saw significant associations of FGT polyps with uterine leiomyoma and 464
endometriosis. Though the cellular composition, characteristics, and mechanism of origin of 465
uterine fibroids and endometrial polyps differ from each other, they can occur concurrently due 466
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
25
to common risk factors like advanced age, obesity, and hormonal dysregulation (Kınay et al., 467
2016; Pavone et al., 2018). Further, increased prevalence of endometriosis among women with 468
FGT polyps emphasises the reported evidence of a higher prevalence of endometrial polyps in 469
endometriosis patients (Shen et al., 2011; Zhang et al., 2018; Lin et al., 2020; Peters et al., 470
2022). 471
Our genetic correlation analysis showed a significant negative correlation between 472
SHBG levels and the risk of FGT polyps. In females, testosterone and oestrogen levels are 473
regulated through their binding to SHBG, which helps maintain hormonal balance in the 474
bloodstream. Lower levels of SHBG can result in elevated levels of free testosterone and free 475
oestrogen. As EPs are oestrogen dependent , lower SHBG may directly stimulate EP 476
development. Moreover, the expression of aromatase 450 enzyme in the EPs, converts free 477
testosterone into oestrogen, further augments the stimulation of polyps' development (Filho et 478
al., 2007) . Thus, our study’s finding of the negative correlation of SHBG in FGT polyps 479
concords with this mechanism. Unexpectedly, our analysis did not reveal a genetic correlation 480
between oestradiol and EPs. The discrepancy between circulating plasma levels and localized 481
endometrial levels of oestradiol may underlie the lack of this correlation. In the literature, serum 482
levels of oestradiol were shown to be significantly lower (five times) than the endometrial 483
oestradiol concentration in the proliferative phase (Huhtinen et al., 2012). The same trend was 484
observed in women with abnormal uterine bleeding and hyperplasia (Cortés-Gallegos et al., 485
1975). Given that abnormal bleeding and hyperplasia are important symptoms of endometrial 486
polyps, it suggests that localised oestradiol may play a more crucial role in EP development. 487
Additionally, a study reported no significant difference in circulating oestradiol concentration 488
among women with and without endometrial polyps (Cortés-Gallegos et al. , 1975) , which 489
further supports our finding. This can also explain the lack of correlation of FGT polyps with 490
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
26
BMI to some extent, as higher BMI is often associated with increased circulating oestradiol 491
levels. 492
To understand how genetic variation impacts trait susceptibility, it is important to link 493
associated genetic variants with specific genes and mechanisms. We employed a diverse range 494
of data layers to map potential candidate genes for the identified genomic loci. Among others, 495
PRIM1 (rs2277339) and COL17A1 (rs193097753) were prioritised since the GWAS signals 496
tagged coding variants in these genes. In the literature, neither of these genes has been directly 497
correlated with FGT development. The PRIM1 gene is a key regulator of DNA replication 498
during cellular proliferation process. It has been shown that the PRIM1 expression was 499
upregulated in breast tumour tissues compared to healthy tissues, and its inhibition led to 500
tumour cell growth regression (Lee et al., 2019). Furthermore, the PRIM1-induced tumour cell 501
growth is stimulated by oestrogen through activation of the oestrogen receptor (ER) (Lee et al., 502
2019). As per one of the postulated hypotheses, polyp development is related to oestrogen 503
stimulation with predominantly increased ER alpha (ER -alpha) and decreased progesterone 504
receptors (PRs) A and B in the glandular epithelium. On the other hand, in the stromal cells of 505
polyps, lower concentrations of ER and PR hinder the decidualization process, preventing them 506
from shedding off during menstruation (Mittal et al., 1996; Peng et al., 2009; Nijkang et al., 507
2019). Since the PRIM1 gene signalling is oestrogen -associated, this can provide important 508
insights into the probable functional mechanisms underlying the development of polyps. The 509
COL17A1 gene is closely associated with epithelial tumour progression and invasiveness (Jones 510
et al. , 2020) , suggesting that a shared biological mechanism related to epithelial cell 511
proliferation could be associated with the development of EPs. However, further functional 512
assays on PRIM1 and COL17A1 genes are needed to confirm these associations. 513
While our study is a large -scale study, it had certain limitations. The cohort from 514
Estonian and Finnish biobanks encompasses women with all kinds of reproductive polyps, and 515
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
27
our phenotype definition did not differentiate between the types/location of polyps to increase 516
study sample size power. However, considering the overall occurrence of polyps in the general 517
population, endometrial polyps would be more prevalent than the other types. A complete 518
understanding of the transcriptome profile of endometrial polyps is still lacking, resulting in 519
knowledge gaps regarding the interplay between genomic loci and the regulation of gene 520
expression. By addressing these knowledge gaps, we can gain valuable insights into the genetic 521
factors and gene expression patterns that contribute to the pathogenesis of endometrial polyps 522
and potentially identify novel therapeutic targets for their management. 523
In conclusion, the first GWAS meta-analysis of FGT polyps highlights and clarifies the 524
genetic mechanisms shared between EP development (tissue overgrowth) and cancerous 525
processes. Furthermore, an analysis of associated diagnoses showed that women diagnosed with 526
EPs have an increased prevalence of other diagnoses, such as endometriosis, uterine fibroids 527
and both benign and malignant neoplasms, which could have implications for patient 528
management and counselling. 529
Supplementary data: Supplementary data are available at Human reproduction online. 530
Data availability: FinnGen cohort level summary statistics can be accessed as described here: 531
https://www.finngen.fi/en/access_results. Protocol for accessing the Estonian Biobank data is 532
described here: https://genomics.ut.ee/en/content/estonian-biobank. GWAS meta -analysis 533
summary statistics will be made available via the GWAS Catalogue (data upload pending). 534
Author’s roles: Authors A.D.S.P., N.P.G, A.S., M.P and T.L. participated in the conception of 535
the study, data analysis, and interpretation of the data, and writing the manuscript. V.R., J.D., 536
and R.M. contributed to data analysis and interpretation. T.L., M.S., M.P., V.R., J.D. and A.S. 537
critically reviewed and provided feedback on each version of the manuscript and all the authors 538
approved the final version. The EstBB Research Team provided the EstBB genotype and 539
phenotype data. 540
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
28
Acknowledgements
We want to acknowledge the participants of the Estonian Biobank and 541
the participants and investigators of the FinnGen study. We also acknowledge the Estonian 542
Biobank Research team members Andres Metspalu, Tonu Esko, Mari Nelis, Georgi Hudjashov, 543
and Lili Milani. 544
Funding: N.P.G. was supported by MATER Marie Sklodowska-Curie which received funding 545
from the European Union’s Horizon 2020 research and innovation programme under grant 546
agreement No. 813707. T.L. was supported by the Estonian Research Council grant PSG776. 547
This study was funded by European Union through the European Regional Development Fund 548
Project No. 2014-2020.4.01.15-0012 GENTRANSMED. Computations were performed in the 549
High-Performance Computing Center of University of Tartu. The study was also supported by 550
the Estonian Research Council (grant no. PRG1076 and MOBJD1056 ) and Horizon 2020 551
innovation grant (ERIN, grant no. EU952516). 552
Conflict of interest: All the authors declared no conflict of interest. 553
554
References
555
556
Altaner S, Gucer F, Tokatli F, Guresci S, Ozdemir C, Puyan FO, Kutlu K. Expression of Bcl-2 557
and Ki -67 in tamoxifen -associated endometrial polyps: comparison with 558
postmenopausal polyps. Onkologie 2006;29:376–380. 559
Annan JJ, Aquilina J, Ball E. The management of endometrial polyps in the 21st century. The 560
Obstetrician & Gynaecologist 2012;14:33–38. 561
Antunes A, Costa-Paiva L, Arthuso M, Costa JV, Pinto -Neto AM. Endometrial polyps in pre-562
and postmenopausal women: factors associated with malignancy. Maturitas 563
2007;57:415–421. 564
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
29
Azevedo JM da R de, Azevedo LM da R de, Freitas F, Wender MCO. Endometrial polyps: 565
when to resect? Arch Gynecol Obstet 2016;293:639–643. 566
Banas T, Pitynski K, Mikos M, Cielecka -Kuszyk J. Endometrial Polyps and Benign 567
Endometrial Hyperplasia Have Increased Prevalence of DNA Fragmentation Factors 40 568
and 45 (DFF40 and DFF45) Together With the Antiapoptotic B -Cell Lymphoma (Bcl-569
2) Protein Compared With Normal Human Endometria. Int J Gynecol Pathol 570
2018;37:431–440. 571
Bateman NW, Dubil EA, Wang G, Hood BL, Oliver JM, Litzi TA, Gist GD, Mitchell DA, 572
Blanton B, Phippen NT, et al. Race-specific molecular alterations correlate with 573
differential outcomes for black and white endometrioid endometrial cancer patients. 574
Cancer 2017;123:4004–4012. 575
Bettocchi S, Achilarre M, Ceci O, Luigi S. Fertility -Enhancing Hysteroscopic Surgery. Semin 576
Reprod Med 2011;29:075–082. 577
Boyle AP, Hong EL, Hariharan M, Cheng Y, Schaub MA, Kasowski M, Karczewski KJ, Park 578
J, Hitz BC, Weng S, et al. Annotation of functional variation in personal genomes using 579
RegulomeDB. Genome Res 2012;22:1790–1797. 580
Cha P-C, Takahashi A, Hosono N, Low S-K, Kamatani N, Kubo M, Nakamura Y. A genome-581
wide association study identifies three loci associated with susceptibility to uterine 582
fibroids. Nat Genet 2011;43:447–450. 583
Chen X, Chen S, Li Y, Gao Y, Huang S, Li H, Zhu Y. SMURF1 -mediated ubiquitination of 584
ARHGAP26 promotes ovarian cancer cell invasion and migration. Exp Mol Med 585
2019;51:1–12. 586
Dal Cin P, Vanni R, Marras S, Moerman P, Kools P, Andria M, Valdes E, Deprest J, Van de 587
Ven W, Van den Berghe H. Four cytogenetic subgroups can be identified in endometrial 588
polyps. Cancer Res 1995;55:1565–1568. 589
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
30
Doria PLS, Moscovitz T, Tcherniakovsky M, Fernandes CE, Pompei LM, Wajman M, 590
Nimwegen AV, Haimovich S. Association of IGF -1 CA(n) and IGFBP3 rs2854746 591
Polymorphisms with Endometrial Polyp Risk. Biomed Res Int 2018;2018:8704346. 592
Dos Santos GA, Viana NI, Pimenta R, Camargo JA de, Guimaraes VR, Romão P, Candido P, 593
Ghazarian V, Reis ST, Leite KRM, et al. Pan-cancer analysis reveals that CTC1-STN1-594
TEN1 (CST) complex may have a key position in oncology. Cancer Genet 2022;262–595
263:80–90. 596
Dreisler E, Stampe Sorensen S, Ibsen PH, Lose G. Prevalence of endometrial polyps and 597
abnormal uterine bleeding in a Danish population aged 20-74 years. Ultrasound Obstet 598
Gynecol 2009;33:102–108. 599
Edwards TL, Hartmann KE, Velez Edwards DR. Variants in BET1L and TNRC6B associate 600
with increasing fibroid volume and fibroid type among European Americans. Hum 601
Genet 2013;132:1361–1369. 602
ENCODE Project Consortium. An integrated encyclopedia of DNA elements in the human 603
genome. Nature 2012;489:57–74. 604
Fatemi HM, Kasius JC, Timmermans A, Disseldorp J van, Fauser BC, Devroey P, Broekmans 605
FJ. Prevalence of unsuspected uterine cavity abnormalities diagnosed by office 606
hysteroscopy prior to in vitro fertilization. Hum Reprod 2010;25:1959–1965. 607
Filho AMB de B, Barbosa IC, Maia H, Genes CC, Coutinho EM. Effects of subdermal implants 608
of estradiol and testosterone on the endometrium of postmenopausal women. Gynecol 609
Endocrinol 2007;23:511–517. 610
Fung JN, Mortlock S, Girling JE, Holdsworth -Carson SJ, Teh WT, Zhu Z, Lukowski SW, 611
McKinnon BD, McRae A, Yang J, et al. Genetic regulation of disease risk and 612
endometrial gene expression highlights potential target genes for endometriosis and 613
polycystic ovarian syndrome. Sci Rep 2018;8:11424. 614
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
31
Ghoussaini M, Mountjoy E, Carmona M, Peat G, Schmidt EM, Hercules A, Fumis L, Miranda 615
A, Carvalho-Silva D, Buniello A, et al. Open Targets Genetics: systematic identification 616
of trait-associated genes using large -scale genetics and functional genomics. Nucleic 617
Acids Res 2021;49:D1311–D1320. 618
GTEx Consortium. The Genotype-Tissue Expression (GTEx) project. Nat Genet 2013;45:580–619
585. 620
Hinckley MD, Milki AA. 1000 office -based hysteroscopies prior to in vitro fertilization: 621
feasibility and findings. JSLS 2004;8:103–107. 622
Humaidan P, Van Vaerenbergh I, Bourgain C, Alsbjerg B, Blockeel C, Schuit F, Van Lommel 623
L, Devroey P, Fatemi H. Endometrial gene expression in the early luteal phase is 624
impacted by mode of triggering final oocyte maturation in recFSH stimulated and GnRH 625
antagonist co-treated IVF cycles. Hum Reprod 2012;27:3259–3272. 626
Jones VA, Patel PM, Gibson FT, Cordova A, Amber KT. The Role of Collagen XVII in Cancer: 627
Squamous Cell Carcinoma and Beyond. Front Oncol 2020;10:352. 628
Karayalcin R, Ozcan S, Moraloglu O, Ozyer S, Mollamahmutoglu L, Batıoglu S. Results of 629
2500 office -based diagnostic hysteroscopies before IVF. Reprod Biomed Online 630
2010;20:689–693. 631
Kho PF, Wang X, Cuéllar-Partida G, Dörk T, Goode EL, Lambrechts D, Scott RJ, Spurdle AB, 632
O’Mara TA, Glubb DM. Multi -tissue transcriptome-wide association study identifies 633
eight candidate genes and tissue -specific gene expression underlying endometrial 634
cancer susceptibility. Commun Biol 2021;4:1211. 635
Kircher M, Witten DM, Jain P, O’Roak BJ, Cooper GM, Shendure J. A general framework for 636
estimating the relative pathogenicity of human genetic variants. Nat Genet 637
2014;46:310–315. 638
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
32
Kınay T, Öztürk Başarır Z, Fırtına Tuncer S, Akpınar F, Kayıkçıoğlu F, Koç S. Prevalence of 639
endometrial polyps coexisting with uterine fibroids and associated factors. Turk J Obstet 640
Gynecol 2016;13:31–36. 641
Koel M, Võsa U, Jõeloo M, Läll K, Gualdo NP, Laivuori H, Lemmelä S; Estonian Biobank 642
Research Team; FinnGen; Daly M , et al. GWAS meta-analyses clarify the genetics of 643
cervical phenotypes and inform risk stratification for cervical cancer. Hum Mol Genet. 644
2023,32:2103-2116. 645
Kurki MI, Karjalainen J, Palta P, Sipilä TP, Kristiansson K, Donner KM, Reeve MP, Laivuori 646
H, Aavikko M, Kaunisto MA, et al. FinnGen provides genetic insights from a well -647
phenotyped isolated population. Nature 2023;613:508–518. 648
Lai T, Qiu H, Si L, Zhen Y, Chu D, Guo R. Long noncoding RNA BMPR1B -AS1 facilitates 649
endometrial cancer cell proliferation and metastasis by sponging miR -7-2-3p to 650
modulate the DCLK1/Akt/NF-κB pathway. Cell Cycle 2022;21:1599–1618. 651
Laisk T, Lepamets M, Koel M, Abner E, Estonian Biobank Research Team, Mägi R. Genome-652
wide association study identifies five risk loci for pernicious anemia. Nat Commun 653
2021;12:3761. 654
Lee SC, Kaunitz AM, Sanchez-Ramos L, Rhatigan RM. The oncogenic potential of endometrial 655
polyps: a systematic review and meta-analysis. Obstet Gynecol 2010;116:1197–1205. 656
Lee W-H, Chen L-C, Lee C-J, Huang C-C, Ho Y-S, Yang P-S, Ho C-T, Chang H-L, Lin I-H, 657
Chang H-W, et al. DNA primase polypeptide 1 (PRIM1) involves in estrogen -induced 658
breast cancer formation through activation of the G2/M cell cycle checkpoint. Int J 659
Cancer 2019;144:615–630. 660
Leeuw CA de, Mooij JM, Heskes T, Posthuma D. MAGMA: generalized gene -set analysis of 661
GWAS data. PLoS Comput Biol 2015;11:e1004219. 662
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
33
Lenci MA, Nascimento VAL do, Grandini AB, Fahmy WM, Depes D de B, Baracat FF, Lopes 663
RGC. Premalignant and malignant lesions in endometrial polyps in patients undergoing 664
hysteroscopic polypectomy. Einstein (Sao Paulo) 2014;12:16–21. 665
Liang H, Cheung LWT, Li J, Ju Z, Yu S, Stemke -Hale K, Dogruluk T, Lu Y, Liu X, Gu C, et 666
al. Whole-exome sequencing combined with functional genomics reveals novel 667
candidate driver cancer genes in endometrial cancer. Genome Res 2012;22:2120–2129. 668
Lin S, Xie X, Guo Y, Zhang H, Liu C, Yi J, Su Y, Deng Q, Zhu W. Clinical characteristics and 669
pregnancy outcomes of infertile patients with endometriosis and endometrial polyps: A 670
retrospective cohort study. Taiwan J Obstet Gynecol 2020;59:916–921. 671
Liu B, Wang T, Jiang J, Li M, Ma W, Wu H, Zhou Q. Association of BET1L and TNRC6B 672
with uterine leiomyoma risk and its relevant clinical features in Han Chinese population. 673
Sci Rep 2018;8:7401. 674
Mittal K, Schwartz L, Goswami S, Demopoulos R. Estrogen and progesterone receptor 675
expression in endometrial polyps. Int J Gynecol Pathol 1996;15:345–348. 676
Nichols CA, Gibson WJ, Brown MS, Kosmicki JA, Busanovich JP, Wei H, Urbanski LM, 677
Curimjee N, Berger AC, Gao GF, et al. Loss of heterozygosity of essential genes 678
represents a widespread class of potential cancer vulnerabilities. Nat Commun 679
2020;11:2517. 680
Nijkang NP, Anderson L, Markham R, Manconi F. Endometrial polyps: Pathogenesis, sequelae 681
and treatment. SAGE Open Med 2019;7:2050312119848247. 682
O’Mara TA, Glubb DM, Amant F, Annibali D, Ashton K, Attia J, Auer PL, Beckmann MW, 683
Black A, Bolla MK, et al. Identification of nine new susceptibility loci for endometrial 684
cancer. Nat Commun 2018;9:3166. 685
Pal L, Niklaus AL, Kim M, Pollack S, Santoro N. Heterogeneity in endometrial expression of 686
aromatase in polyp-bearing uteri. Hum Reprod 2008;23:80–84. 687
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
34
Pavone D, Clemenza S, Sorbi F, Fambrini M, Petraglia F. Epidemiology and Risk Factors of 688
Uterine Fibroids. Best Pract Res Clin Obstet Gynaecol 2018;46:3–11. 689
Peng X, Li T, Xia E, Xia C, Liu Y, Yu D. A comparison of oestrogen receptor and progesterone 690
receptor expression in endometrial polyps and endometrium of premenopausal women. 691
J Obstet Gynaecol 2009;29:340–346. 692
Peters M, Mikeltadze I, Karro H, Saare M, Estonian Biobank Research Team, Salumets A, 693
Mägi R, Laisk T. Endometriosis and irritable bowel syndrome: similarities and 694
differences in the spectrum of comorbidities. Hum Reprod 2022;37:2186–2196. 695
Pujol-Gualdo N, Läll K, Lepamets M, Estonian Biobank Research Team, Rossi H-R, Arffman 696
RK, Piltonen TT, Mägi R, Laisk T. Advancing our understanding of genetic risk factors 697
and potential personalized strategies for pelvic organ prolapse. Nat Commun 698
2022;13:3584. 699
Rahmioglu N, Mortlock S, Ghiasi M, Møller PL, Stefansdottir L, Galarneau G, Turman C, 700
Danning R, Law MH, Sapkota Y, et al. The genetic basis of endometriosis and 701
comorbidity with other pain and inflammatory conditions. Nat Genet 2023;55:423–436. 702
Raz N, Feinmesser L, Moore O, Haimovich S. Endometrial polyps: diagnosis and treatment 703
options - a review of literature. Minim Invasive Ther Allied Technol 2021;30:278–287. 704
Roadmap Epigenomics Consortium, Kundaje A, Meuleman W, Ernst J, Bilenky M, Yen A, 705
Heravi-Moussavi A, Kheradpour P, Zhang Z, Wang J, et al. Integrative analysis of 111 706
Reference
human epigenomes. Nature 2015;518:317–330. 707
Sahoo SS, Aguilar M, Xu Y, Lucas E, Miller V, Chen H, Zheng W, Cuevas IC, Li H-D, Hitrys 708
D, et al. Endometrial polyps are non -neoplastic but harbor epithelial mutations in 709
endometrial cancer drivers at low allelic frequencies. Mod Pathol 2022;35:1702–1712. 710
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
35
Sasaki LMP, Andrade KRC, Figueiredo ACMG, Wanderley M da S, Pereira MG. Factors 711
Associated with Malignancy in Hysteroscopically Resected Endometrial Polyps: A 712
Systematic Review and Meta-Analysis. J Minim Invasive Gynecol 2018;25:777–785. 713
Schmitt AD, Hu M, Jung I, Xu Z, Qiu Y, Tan CL, Li Y, Lin S, Lin Y, Barr CL, et al. A 714
Compendium of Chromatin Contact Maps Reveals Spatially Active Regions in the 715
Human Genome. Cell Rep 2016;17:2042–2059. 716
She J, Su D, Diao R, Wang L. A Joint Model of Random Forest and Artificial Neural Network 717
for the Diagnosis of Endometriosis. Front Genet 2022;13:848116. 718
Shen L, Wang Q, Huang W, Wang Q, Yuan Q, Huang Y, Lei H. High prevalence of endometrial 719
polyps in endometriosis-associated infertility. Fertil Steril 2011;95:2722-2724.e1. 720
Shiratori A, Okumura K, Nogami M, Taguchi H, Onozaki T, Inoue T, Ando T, Shibata T, Izumi 721
M, Miyazawa H. Assignment of the 49 -kDa (PRIM1) and 58 -kDa (PRIM2A and 722
PRIM2B) subunit genes of the human DNA primase to chromosome bands 1q44 and 723
6p11.1-p12. Genomics 1995;28:350–353. 724
Takeda T, Banno K, Kobayashi Y, Adachi M, Yanokura M, Tominaga E, Kosaki K, Aoki D. 725
Mutations of RAS genes in endometrial polyps. Oncol Rep 2019;42:2303–2308. 726
Tanos V, Berry KE, Seikkula J, Abi Raad E, Stavroulis A, Sleiman Z, Campo R, Gordts S. The 727
management of polyps in female reproductive organs. Int J Surg 2017;43:7–16. 728
Uglietti A, Buggio L, Farella M, Chiaffarino F, Dridi D, Vercellini P, Parazzini F. The risk of 729
malignancy in uterine polyps: A systematic review and meta -analysis. Eur J Obstet 730
Gynecol Reprod Biol 2019;237:48–56. 731
Unler GK, Gokturk HS, Toprak E, Erinanc OH, Korkmaz H. Does the Presence of Endometrial 732
Polyp Predict Colorectal Polyp? Am J Med Sci 2016;351:129–132. 733
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
36
Vieira MDC, Vitagliano A, Rossette MC, Neto LC de A, Gallo A, Sardo ADS. Endometrial 734
Polyps: Update Overview on Etiology, Diagnosis, Natural History and Treatment. 735
CEOG 2022;49:232. IMR Press. 736
Vitale SG, Haimovich S, Laganà AS, Alonso L, Di Spiezio Sardo A, Carugno J, From the 737
Global Community of Hysteroscopy Guidelines Committee. Endometrial polyps. An 738
evidence-based diagnosis and management guide. Eur J Obstet Gynecol Reprod Biol 739
2021;260:70–77. 740
Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from 741
high-throughput sequencing data. Nucleic Acids Res 2010;38:e164. 742
Watanabe K, Taskesen E, Bochoven A van, Posthuma D. Functional mapping and annotation 743
of genetic associations with FUMA. Nat Commun 2017;8:1826. 744
Wu L, Wang Y, Liu Y, Yu S, Xie H, Shi X, Qin S, Ma F, Tan TZ, Thiery JP, et al. A central 745
role for TRPS1 in the control of cell cycle and cancer development. Oncotarget 746
2014;5:7677–7690. 747
Yi SE, LaPolt PS, Yoon BS, Chen JY, Lu JK, Lyons KM. The type I BMP receptor BmprIB is 748
essential for female reproductive function. Proc Natl Acad Sci U S A 2001;98:7994–749
7999. 750
Zhang Y-N, Zhang Y-S, Yu Q, Guo Z -Z, Ma J -L, Yan L. Higher Prevalence of Endometrial 751
Polyps in Infertile Patients with Endometriosis. Gynecol Obstet Invest 2018;83:558–752
563. 753
754
755
756
757
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
37
Supplementary materials: 758
Supplementary figure 1: Genome-wide gene association analysis of our GWAS dataset 759
using ‘Multi-marker Analysis of GenoMic Annotation (MAGMA) wherein input SNPs 760
were mapped to 18895 protein -coding genes. On the Manhattan plot, the y -axis represents 761
−log10(P-values) for the association of variants identified in polyps of the female genital tract. 762
The horizontal red dashed line represents the genome -wide significance threshold (P < 5 × 763
10−8) and prioritised genes for each locus are shown. Genes associated with exonic missense 764
variants are denoted in red text. 765
766
Supplementary Tables: 767
ST 1 Genes mapped by expression quantitative trait loci (eQTL) and Chromatin interactions 768
by Functional Mapping and Annotation of Genome-Wide Association Studies (FUMA 769
GWAS) 770
ST 2 Results of the Multi -marker Analysis of GenoMic Annotation (MAGMA) gene -based 771
analysis 772
ST 3 Results of the GWAS catalogue lookup 773
ST 4 Look-up of the 10 lead variants in endometrial cancer summary statistics 774
(https://pubmed.ncbi.nlm.nih.gov/31040137/) 775
ST 5 Genetic correlation between female genital tract polyps and other traits 776
ST 6 Phenotypes associated with Polyps of the female genital tract 777
ST 7 Results of survival analysis lookup in FinnGen data 778
779
780
781
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint
The copyright holder for thisthis version posted January 30, 2024. ; https://doi.org/10.1101/2024.01.29.24301773doi: medRxiv preprint
Text is read by the "Ask this paper" AI Q&A widget below.
Extraction quality varies by source — PMC NXML preserves structure
cleanly, OA-HTML may include some navigation residue, and OA-PDF can
have broken hyphenation. The publisher copy
(via DOI)
is the canonical version.