This article is mentioned in:

Introduction

Endometriosis is an enigmatic, benign, hormone-dependent gynecological disorder, defined by the presence of functional endometrium-like tissue (endometrial glands and stroma) external to the uterine cavity and affecting ~10% of women of reproductive age. This condition markedly affects the health of women, as well as their quality of life (1). Its main clinical symptoms include chronic pelvic pain, dyspareunia, dysmenorrhea and impaired fertility, with laparoscopy and biopsy remaining the gold-standard methods for its diagnosis (2).

Despite significant progress being made in the understanding of the pathophysiology of endometriosis and the various theories suggested thus far, its precise etiology remains unknown. The disease exhibits high complexity, as genetic, epigenetic and environmental factors appear to interact in order to formulate the disease phenotype. Furthermore, this multifactorial nature of endometriosis involves the participation of aberrant immunological responses, angiogenesis processes and biochemical alterations. Of note, endometriosis has a strong genetic predisposition (3,4), as shown by studies conducted on monozygotic twins and families (5), gene association studies that took into account candidate genes and single nucleotide polymorphisms (SNPs) (6), as well as genome-wide association studies (GWAS) (7). For example, eight GWAS of women of European and East Asian origin that have been published to date (2) have identified 19 distinct disease-associated signals harbored at 14 loci (8). Moreover, a meta-analysis of 15 GWAS and a new replication analysis, including 58,115 cases and 733,480 controls in total, revealed 27 genetic loci associated with endometriosis at the genome-wide P-value threshold (P<5x10-8), 13 of which are novel, while an additional 8 novel genes identified from gene-based association analyses (9).

The evidence of an association among genetic polymorphisms and the risk of developing endometriosis is robust (10). A number of studies have reported a nine-fold increase in the risk of developing endometriosis among women from the East Asian population compared with the European or American women populations (4,11,12). Although the risk of endometriosis has been strongly linked to ethnicity, the main differences identified in population groups have not been well defined. What remains to be understood is how risk variation is linked to ethnicity and the minor differences in SNP variation, as captured in the recent geographical and evolutionary history of humans. Of note, differences in autoimmune disease risk variants have also been reported across different continental populations. In particular, various data have demonstrated that several systemic lupus erythematosus susceptibility loci confer risk across multiple ethnicities, while distinct differences in the risk of developing the disease were found when populations of European, African, African-American, Asian and Hispanic ancestry were analyzed (13).

In Africa, ~45,000 to 60,000 years ago, a marked demographic and geographic spread began that rapidly brought human presence to almost all habitable areas of the earth. Genetic and paleoanthropological studies are consistent with the aforementioned evidence. Genomic data from modern humans suggest that this expansion was accompanied by a continuous loss of genetic diversity, a result of what is known as the ‘serial founder effect’ (14). It is generally assumed that the bottleneck occurred as a small group(s) with an effective population size of only ~2,000 individuals migrated from the African continent to the Near East (14-16). During the great expansion, there was an uninterrupted and considerable reduction in genetic diversity proportional to the geographic distance from the African homeland, which is indicated by the motif of average heterozygosities of contemporary populations (14). However, genomes from substructured populations retain a numerous amount of unique variants. As a result of the relatively profound substructure within the continent, genetic variation in Africa varies considerably from region to region. Groups such as the Khoisan, Hadza, Sandawe and Forest Pygmies have been shown to maintain extremely high genetic diversity, relative to out-of-Africa populations, as evidenced by studies on autosomal DNA polymorphism patterns in present-day African hunter-gatherers (14-20).

Since nowadays the available genetics data (e.g. from NGS, GWAS) are in abundance, there is an opportunity for drawing conclusions, from screening populations for medical genetic studies, for mapping genotypes to phenotypes and for rendering the power of natural selection in human history.

However, the gathering of huge data impedes the identification of such genetic patterns in order to obtain SNPs as genetic markers for a number of applications such pharmacogenomics, personalized medicine and genetics. This task requires the implementation of thousands of SNPs, which are associated with various clinical disorders in correlation studies. In the present study, an attempt was made to examine a wide set of genetic targets that characterize endometriosis through the prism of the genetic profile of geographically distributed population groups and the allele frequency of targeted SNPs. To this end, bioinformatics studies, such as those presented herein, that join the world-wide genetic information represented by ‘1000 Genomes Project’ with a disease-related SNP database such as Demetra (21) may alleviate the aforementioned issue, providing a simple and efficient handling of vast information and this could lead to the export of results.

Data and methods

Genetic targets dataset

Endometriosis-related SNPs were extracted from the Demetra Application database (DAD) (21) using the Demetra Application webserver and have been used in the pipeline of the present study. The extracted SNPs were labeled with the same classification of the DAD and contained accompanying information, including the SNP identification number based on the dbSNP database (22), chromosome name, genetic locus, SNP type (introns or exons), the name of the responsible gene or the name of the intragenic region, and the classification based on the DAD (three different classes, ‘strongly-associated SNPs’, ‘highly-associated SNPs’ and ‘associated SNPs’ with endometriosis).

Population studies

In the present study, five geographical population groups were studied, focusing on Europeans, Africans, Americans, East Asians and South Asians. Sample sizes and origins of the individual population samples are provided in the International Genome Sample Resource (IGSR) (23), which have been developed under the ‘1000 Genomes Project’ (24). The present study was performed using human genomes that were contained in the phase three collection of the IGSR on reference assembly GRCh38 (25,26). Since the ‘1000 Genomes Project’ has created call sets of sequence variants for each of the different genomes sequenced, the downloaded data were in multi-individual variant call format (VCF) (27) per chromosome, with genotypes listed for each sample (26). The MATLAB Bioinformatics toolbox (28) was used for pre-analyzing the extracted data (VCF format) and storing them in a structured database per chromosome and genetic locus of the identified variants, and per origin, including using the IGSR directory of the individuals.

SNPs analysis and sensitive targets

The Python programming language was used to identify the endometriosis-related SNPs from the DAD, in the structured database of variants that was developed from the IGSR data. The selected variants of SNPs were stored in a newly structured database by combining the information from both of the aforementioned databases. Each entry in the first columns contains the information from the DAD and the following columns then contain the information for the VCF file for each individual. Subsequently, all the identified variants were pre-analyzed to collect the allele frequencies for all SNPs, as estimated in the IGSR. Specifically, for each SNP, information was collected regarding the allele frequency of appearances in Europeans, Africans, Americans, East Asians and South Asians.

A specified analysis was performed for drawing the disease genomic ‘grammar’ (DGG) profiles of endometriosis in the five major groups of ethnic origins. All the variants were studied as candidate genetics targets for each group using the Python programming language. An algorithm was designed to be able to detect all the allele frequency ‘alerts’ in each studied group. The algorithm was able to classify each allele frequency in three clusters including a) ‘Low’, SNP allele frequency ≤0.1; b) ‘Normal’, 0.1 < SNP allele frequency <0.9; and c) ‘High’, SNP allele frequency ≥0.9. For the purposes of the present study, ‘alerts’ were issued only for cases when the allele frequencies belong to the ‘Low’ or ‘High’ clusters. Subsequently, by using specific R programming language packages, such as ‘BioCircos’ (available online at http://bioinfo.ibp.ac.cn/biocircos/), ‘circlize’ (available at http://cran.r-project.org/web/packages/circlize/) and ‘CMplot’ (available at https://github.com/xiaolei-lab/rMVP), the results were analyzed to draw the circular visualization and the result visualization (29-31).

Results

Endometriosis-related genetic targets and the ‘1000 Genomes Project’

The DAD (21), an integrated database with all SNPs associated with endometriosis, was used in the pipeline of the present study. The DAD holds information on 4,772 SNPs corresponding to genes, alleles, pseudogenes, transcription factors and intergenic regions. From the 4,772 SNPs, 68 SNPs were identified in gene-coding region sites (1.5%) and the remaining (non-coding) corresponded either to non-translated regulatory regions or to intragenic regions (98.5%) (Fig. 1A). Based on the results from the Demetra Application, the endometriosis-related SNPs were classified into three major classes, including ‘strongly-associated SNPs’ with 50 members, ‘highly-associated SNPs’ with 125 members and ‘associated SNPs’ with 4,600 members (Fig. 1B). The ‘1000 Genomes Project’ phase 3 has systematically completed mapping the genomes of >2,500 individuals for genetic variation (32) of the five major groups, including Europeans, Africans, Americans, East Asians and South Asians (33,34).

Figure 1 Based on the content of the Demetra Application database for endometriosis-related SNPs, the present study included sequence variants in (A) coding and non-coding regions, which (B) in total, could be classified into three major classes, based on the rate of connection with this specific disease. SNP, single nucleotide polymorphism.

A clear separation of the DGG of endometriosis is shown in the allele frequencies, as identified in the five major clusters of the human population. The incidence of different genetic variants in a population, represented by their allele frequency, is a reflection of the genetic diversity, and changes in allele frequencies over time can indicate that genetic drift is occurring or that new significant variants have been introduced into the population (34-37). From the ‘1000 Genomes Project’ database (32), each identified variant is described by a specific allele frequency, which is estimated by dividing the number of times the allele of interest is observed in the population by the total number of incidence of all the alleles present for that particular genetic locus in the population (38,39). The ‘1000 Genomes Project’ provides the general allele frequency of the identified variants and the corresponding allele frequencies of the five major groups, including Europeans, Africans, Americans, East Asians and South Asians (33,34). Allele frequencies are presented as a decimal in the range of 0-1.

Directional change and reversal in allele frequencies has been shown in the 4,775 endometriosis-related SNPs between the individuals of the major five groups. The histogram analysis of endometriosis-related SNPs using the ‘1000 Genomes Project’ dataset revealed clearly different distributions in the five major geographical population groups (Fig. 2). The majority of the groups exhibited a multimodal distribution, since several peaks are close together and the top of the distribution forms a plateau (40). Although different allele frequencies were identified in the studied SNPs, some groups appeared to have similar distributions, but with different quantifications, as in between the Africans and East Asians or the Europeans and the Americans (Fig. 2). The five studied geographical groups accumulated different SNP totals at the sensitive two ends of the distribution of the allele frequencies, including the cluster of the ‘low allele frequencies’ (SNP allele frequency ≤0.1) and the cluster of the ‘high allele frequencies’ (SNP allele frequency ≥0.9) (Figs. 3A and 4A) (41-43).

Figure 2 Histogram of the allele frequencies of endometriosis-related SNPs as extracted from genome-wide association studies data from the ‘1000 Genomes Project’. (A) Histogram of the SNP allele frequencies for the separated group of Europeans. (B) Histogram of the SNP allele frequencies for the separated group of Africans. (C) Histogram of the SNP allele frequencies for the separated group of Americans. (D) Histogram of the SNP allele frequencies for the separated group of East Asians. (E) Histogram of the SNP allele frequencies for the separated group of South Asians. (F) Histogram of the SNP allele frequencies for total number of studied individuals. SNP, single nucleotide polymorphism; EUR, Europeans; AFR, Africans; AMR, Americans; EAS, East Asian; SAS, South Asians.

Figure 3 (A) Circos-like visualizations of the genomic data per chromosome of the five major population groups in the identified Endometriosis related SNPs with low allele frequencies (blue dots, Europeans; green dots, Africans; yellow dots, Americans; brown dots, East Asians; red dots, South Asians). (B) Venn diagram of the five major populations identified endometriosis-related SNPs with low allele frequencies (populations are labeled with ‘A’ for Europeans, ‘B’ for Africans, ‘C’ for Americans, ‘D’ for East Asians, ‘E’ for South Asians and their colors are the same as those in panel A). SNP, single nucleotide polymorphism.

Figure 4 (A) Circos-like visualizations of the genomic data per chromosome of the five major population groups in the identified endometriosis-related SNPs with high allele frequencies (blue dots, Europeans; green dots, Africans; yellow dots, Americans; brown dots, East Asians; red dots, South Asians). (B) Venn diagram of the five major geographical populations identified endometriosis-related SNPs with high allele frequencies (populations are labeled with ‘A’ for Europeans, ‘B’ for Africans, ‘C’ for Americans, ‘D’ for East Asians, ‘E’ for South Asians and their colors are the same as those in panel A). SNP, single nucleotide polymorphism.

Different totals and reference SNPs were accumulated in the low and high clusters among the different population groups (Figs. 3 and 4). The East Asian group had the largest sample of SNPs with low allele frequencies, followed by the Africans, South Asians, Americans and Europeans (Fig. 3). On the other hand, the African group had the largest sample of different SNPs as expected (44) with high allele frequencies followed by the East Asian group (Fig. 4). The South Asian, American and European groups exhibited markedly fewer totals in SNPs with high allele frequencies (Fig. 4). Although some population groups shared some similarities in the identified SNPs in the low and high clusters, the overall distribution of the endometriosis-related SNPs and their genetic locus per chromosome in the five studied population groups revealed clear differentiation (Figs. 3A and 4A). All the population groups shared a common genetic background in 296 SNPs with low allele frequencies and six SNPs with high allele frequencies (Figs. 3B and 4B). These results may be associated with clinical data in order to understand common or population specific aspects in endometriosis and form a basis for research towards understanding the genetic components of the disease that may aid in improving early diagnosis or directed medical treatment for patients (45,46).

Key genetic targets between the five major population groups in endometriosis

Allele frequencies in endometriosis-related SNPs revealed different genetic diversity in the studied geographical population groups. The genetic SNP distribution for endometriosis begins from the African population, and follows the Homo sapiens expansion history, beginning with Africa (47). This scenario is also reflected in the data of the present study, since in the African population genomes from the substructured populations of Africa retain an exceptional number of unique variants. The majority of the endometriosis-related SNPs with low and high allele frequencies were identified (Tables I and II) in Africa. However due to the founder's effect, there is a loss of genetic diversity in the selected population groups on endometriosis-associated SNPs (48). Major common and different endometriosis-related genetic targets were identified in the five population groups using the African population group results as a basis. As Kobayashi et al (49) mentioned, the number of SNPs detected at a low allele frequency was much higher than the number of SNPs detected at a higher allele frequency.

Table I Key genetic targets identified in endometriosis-related SNPs with low allele frequencies.

Table I

Key genetic targets identified in endometriosis-related SNPs with low allele frequencies.

A/A	Class-dbSNP ID-gene	Europeans	Africans	Americans	East Asians	South Asians
1	A - rs144240142 # MAP3K4	√	√	√	√	√
2	A - rs2479037 # VTI1A	√		√	√	√
3	A - rs11549465 # HIF1A	√	√	√	√
4	A - rs700519 # CYP19A1	√			√
5	A - rs1048943 # CYP1A1	√	√	√
6	A - rs4072111 # IL16	√	√
7	A - rs11556218 # IL16	√				√
8	A - rs10928050 # KAZN		√			√
9	A - rs2235529 # WNT4		√
10	A - rs20417 # PACERR				√
11	A - rs1056836 # CYP1B1				√
12	A - rs4064 # FAS				√
13	B - rs14647 # NSD2	√	√	√	√	√
14	B - rs13177597 # ATP6AP1L	√	√	√	√	√
15	B - rs1800796 # IL6-AS1	√	√
16	B - rs74491657 # 7p12.3	√	√
17	B - rs4729645 # MUC17	√	√	√
18	B - rs74974199 # MUC17	√	√	√	√	√
19	B - rs113759408 # CYP11B1	√		√	√	√
20	B -rs146597587 # IL33	√	√	√	√	√
21	B - rs10965235 # CDKN2B-AS1	√		√		√
22	B - rs11245936 # MUC2	√	√	√	√	√
23	B - rs7103978 # MUC2	√	√	√	√	√
24	B - rs11245954 # MUC2	√		√	√	√
25	B - rs61764370 # KRAS	√	√	√	√	√
26	B - rs76731691 # CEP112	√	√	√	√	√
27	B - rs34536443 # TYK2	√	√	√	√	√
28	B - rs12720356 # TYK2	√	√	√	√	√
29	B - rs280523 # TYK2	√		√	√	√
30	B - rs2304256 # TYK2		√
31	B - rs12720270 # TYK2		√
32	B - rs12038474 # CDC42		√
33	B - rs10917151 # CDC42		√
34	B - rs7412010 # CDC42/WNT4		√
35	B - rs12037376 # WNT4		√
36	B - rs61768001 # WNT4		√
37	B - rs3820282 # WNT4		√
38	B - rs56318008 # WNT4		√
39	B - rs55938609 # WNT4		√
40	B - rs2510770 # PDLIM5		√
41	B - rs246832 # MMP2		√	√	√	√
42	B - rs2226158 # LINC02523		√
43	B - rs71575922 # SYNE1		√	√	√	√
44	B - rs62468795 # RPS2P32		√	√	√	√
45	B - rs1042839 # PGR		√		√	√
46	B - rs74485684 # LOC105376607		√		√	√
47	B - rs10835638 # FSHB		√		√	√
48	B - rs11031047 # LOC105376607		√		√	√
49	B - rs174538 # FEN1		√
50	B - rs17727841 # IGF1		√	√
51	B - rs4783689 # CDH1		√
52	B - rs1799969 # ICAM1		√		√	√
53	B - rs2856836 # IL1A				√
54	B - rs1304037 # IL1A				√
55	B - rs17561 # IL1A				√
56	B - rs9322331 # ESR1				√
57	B - rs1544410 # VDR				√
58	B - rs11592737 # CYP2C19				√
59	B - rs1802669 # MLLT10				√

[i] The five major population groups studied were Europeans, Africans, Americans, East Asians and South Asians. Class A is the strongly-associated SNPs and class B is the highly-associated SNPs with endometriosis. SNP, single nucleotide polymorphism.

Table II Key genetic targets identified in endometriosis-related SNPs with high allele frequencies.

Table II

Key genetic targets identified in endometriosis-related SNPs with high allele frequencies.

A/A	Class-dbSNP ID-gene	Europeans	Africans	Americans	East Asians	South Asians
1	A - rs2258447 # MUC4	√		√
2	A - rs2427284 # LAMA5	√		√	√
3	A - rs13394619 # GREB1		√
4	A - rs2688513 # MUC4		√
5	A - rs1570360 # VEGFA		√
6	A - rs244285 # STXBP4		√
7	A - rs383369 # LILRB2		√
8	A - rs1250248 # FN1				√
9	A - rs9582036 # FLT1					√
10	B - rs1209731 # NME7	√	√	√	√	√
11	B - rs1894692 # SLC19A2 / F5	√	√	√	√	√
12	B - rs495590 # DNM3		√
13	B - rs1250241 # FN1	√	√		√
14	B - rs1250247 # FN1		√		√
15	B - rs7739264 # LOC100506885		√
16	B - rs796945 # RNLS		√
17	B - rs2856111 # MUC2			√

[i] The five major population groups studied were Europeans, Africans, Americans, East Asians and South Asians. Class A is the strongly-associated SNPs and class B is the highly-associated SNPs with endometriosis.

The genomic ‘grammar’ of endometriosis has 296 common genetic targets of SNPs with low allele frequencies and six common genetic targets of SNPs with high allele frequencies in the five studied population groups (Figs. 3 and 4). Within the Class A and Class B (strongly- and highly-associated SNPs with endometriosis) 11 key genetic targets with a low allele frequency were identified, including SNPs of the MAP3K4, NSD2, MUC2, MUC17, IL33, KRAS and CEP112 genes (Table I) (50-57), as well as two important genetic targets with a high allele frequency, including the SNPs of the NME7 gene (Table II) (58-61). A significant differentiation has been identified in SNPs of the Africans and East Asians populations (62,63). The African population has 23 unique genetic targets, of which 15 were identified with a low allele frequency and correspond to the WNT4, TYK2, CDC42, PDLIM5, FEN1 and CDΗ1 genes (56,64-68), and eight were identified with a high allele frequency and correspond to the GREB1, MUC4, VEGFA, STXBP4, LILRB2, DNM3 and RNLS genes (44,69-76) (Tables I and II). The East Asian population had 11 unique genetic targets, of which ten were identified with a low allele frequency and correspond to the PACERR, CYP1B1, FAS, IL1A, ESR, VDR, CYP2C19 and MLLT10 genes (8,77-86), and one was identified with a high allele frequency and corresponds to the FN1 gene (8) (Tables I and II). The European population exhibited a notable differentiation in the IL16(87), MUC4(88) and LAMA5(88) genetic targets, from which MUC4 also corresponds to the American population and LAMA5 to the American and East Asian populations (10,88,89). This is in agreement with the general observation that the genetic distance between Asia and Africa is shorter than that between Africa and the other continents, as both Africans and Asians contributed to the settlement of Europe (90).

The genetic profile of endometriosis differs between the five major geographical population groups

The differentiation of endometriosis based on its genomic ‘grammar’ among population groups is not a hypothetical scenario. The allele frequency model of the target SNPs related to endometriosis may explain the degree of differentiation, depending on the initial genetic background of each population group (Table I, Table II and Table III and Fig. 5). An even more notable finding is that this differentiation appears to be consistent with the continuous loss of genetic diversity along the geographical expansion of Homo sapiens on earth, and the way that they have conquered the continents (Fig. 5) (91,92). The African population appears to have the widest genomic profile for the manifestation of endometriosis followed by the genomic profile of East Asians, Europeans, Americans and South Asians (Fig. 5 and Table III). Almost all genetic variants that affect the risk of developing endometriosis have human-specific origins, with ancient roots that often trace back to the Africans population (93,94). It is currently believed by scientists that SNPs, biological pathways and consequently systems that are involved in several diseases may have more ancient origins and each of these ancient variations generated the conditions for modern disease (94). Indicatively, endometriosis is a disease that has been recognized in other mammals, including the olive baboon (Papio anubis) and guinea pig (Cavia tschudii) (95,96).

Figure 5 Schematic representation of the genomic ‘grammar’ of endometriosis in the five major groups of populations based on the endometriosis-related single nucleotide polymorphism classification groups and the two allele frequencies clusters (low allele frequencies cluster colored in blue gradient and high allele frequencies cluster colored in red gradient, beside each group. Europeans are indicated in blue color, Africans in green, Americans in yellow, East Asians in brown, and South Asians in red color). The association of populations groups with the genomic ‘grammar’ of endometriosis is presented by the length of colored arc, beside each population group. The arrows indicate the potential migration paths (A-G) in the pattern of human migration from Africa.

Table III Percentage of the identified SNPs associated with endometriosis (clusters A-C) in the different origin groups including Europeans, Africans, Americans, East Asians and South Asians.

Table III

Percentage of the identified SNPs associated with endometriosis (clusters A-C) in the different origin groups including Europeans, Africans, Americans, East Asians and South Asians.

A/A	Category	Acronym	Europeans (%)	Africans (%)	Americans (%)	East Asians (%)	South Asians (%)
1	Strongly-associated SNPs (low)	A Low	13.70	11.75	7.84	13.72	7.84
2	Highly-associated SNPs (low)	B Low	15.90	31.85	16.81	24.77	19.46
3	Associated SNPs (low)	C Low	14.01	24.60	17.27	24.88	18.84
4	Strongly-associated SNPs (high)	A High	3.92	9.80	3.92	3.92	1.96
5	Highly-associated SNPs (high)	B High	1.70	6.19	2.65	3.53	1.76
6	Associated SNPs (high)	C High	0.41	9.24	1.26	7.08	1.88

Human populations exhibit differences in the allele frequency of several common and rare SNPs. These observations are largely the effect of the diverse environmental, cultural, demographic and genomic scenarios of modern human populations. High rates of endometriosis in East Asian and European populations have been previously reported (4,97,98). However, to the best of our knowledge, the strong association of endometriosis with African populations through a specific set of key genetic targets is presented for the first time herein (Table I, Table II and Table III and Fig. 5), although studies have begun to converge in this direction (99,100). This may be due to the lack of reported cases of endometriosis in the African population (63,101). On the other hand, it should be considered that African women may experience different epigenetic factors that act negatively in addition to their burdened genomic profile for the onset of the disease. Apart from the strong link to hereditary factors, environmental exposures can result in the development of endometriosis. An increased risk of developing endometriosis is associated with plethora of environmental factors, such as persistent organochlorine pollutants, perfluorochemicals, elevated levels of phthalate esters and exposure to cigarette smoke (98).

Discussion

A set of 13 key genetic targets have been identified in the genomic ‘grammar’ of endometriosis and are representative of all the studied populations groups. Based on the findings of the present study, eleven were identified in low allele frequencies and two in high allele frequencies (Tables I and II). Moreover, nine were identified in gene loci and four in epigenetic-related loci. A literature review revealed three confirmed genes through genome-wide genetic analyses, including MAP3K4 (rs144240142) (50), NSD2 (rs14647) (57) and IL33 (rs146597587) SNPs (102,103). It is notable that four polymorphisms do not involve genes, but epigenetic-associated targets with endometriosis, including CEP112 (rs76731691) (61), NME7 (rs1209731) (61), Factor V Leiden (F5, rs1894692) (61) and the unnamed genetic locus for the rs13177597(61) SNPs. Indicatively, the polymorphism rs1894692, an intergenic region between SLC19A2 and F5, has been proven to be a critical epigenetic target through a GWAS (104) and has been directly linked to endometriosis (86). The polymorphisms involving the MUC17 (rs74974199) (52), MUC2 (rs11245936, rs7103978) (21,51), KRAS (rs61764370) (85,105) and TYK2 (rs34536443, rs12720356) (106-108) genes, although they have been confirmed to be associated with endometriosis in separated subgroups of the population, are proposed for study in genome-wide genetic analyses based on the results of the present study.

A total of 36 polymorphisms involving 35 genes and one epigenetic target [rs7412010(109)], were found to markedly differ in high and low allele frequencies, and appear to be representative in one of the studied populations (23 in the African population, 11 in the East Asian population, one in the South Asian population and one in the American population, Tables I and II). All the other polymorphisms were identified in more than one population groups. The genomic ‘grammar’ of endometriosis in Africans and East Asians markedly differs and contains the majority of the SNPs in the two sensitive clusters (low and high) than the other population groups (Figs. 3B and 4B). The African population has a markedly different number of sensitive polymorphisms related to endometriosis in low and high allele frequencies, including the genes TYK2 (rs2304256 and rs12720270), WNT4 (rs2235529, rs12037376, rs61768001, rs3820282, rs56318008 and rs55938609), CDC42 (rs12038474 and rs10917151), PDLIM5 (rs2510770), VEGFA (rs1570360), LILRB2 (rs383369), CDH1 (rs4783689) and GREB1 (rs13394619) (56). GWAS performed confirm these findings for the genes TYK2 (110,111) LILRB2(112) and CDC42 (56,113). The study performed by Välimäki et al (65) demonstrated the more frequent occurrence of uterine leiomyoma (UL) in women of African origin, and suggested the WNT4 as a candidate predisposition gene that may play a critical role in uterine pathology and in particular endometriosis. However only two sensitive polymorphisms of the WNT4 gene have been studied, including rs3820282 and rs3820282. Although CDH1, PDLIM5, VEGFA, LILRB2, SYNE1 and GREB1 gene-related polymorphisms are directly associated with endometriosis and endometrial cancer, to the best of our knowledge, they have never been studied as sensitive key genetic targets in women with endometriosis in the African population (61,72,114). The specialized genomic ‘grammar’ of endometriosis in East Asians includes gene-related polymorphisms in low and high allele frequencies, including PACERR (rs20417), CYP1B1 (rs1056836), FAS (rs4064), IL1A (rs2856836, rs1304037, rs17561), ESR1 (rs9322331), VDR (rs1544410), CYP2C19 (rs11592737), MLLT10 (rs1802669) and FN1 (rs1250248). GWAS performed confirm these findings for the genes FN1(7), MLLT10(9), CYP2C19(115), IL1A (82) and CYP1B1(78). However, the reported SNPs regarding the VDR, ESR1, FAS and PACERR gene loci have never been studied as sensitive key genetic targets in women with endometriosis in the East Asian population, at least to the best of our knowledge.

The present study revealed notable ‘key’ genetic targets in the genomic ‘grammar’ of endometriosis, with the evidence of population-based heterogeneity. Several of the results obtained herein appear to be confirmed through a number of scientific publications. On the other hand, some other ‘key’ genetic targets which are presented herein may be worthy of further investigations in order to understand the specialized endometriosis genomic ‘grammar’ of each population group in the future. Moreover, further analysis is required to indicate how the common genes within the population groups presented in the present study, or the unique genes in a specific group, are involved in one or more biological pathways, resulting in disease development, aiming to provide personalized medicine and preventive measures for different population groups.

In conclusion, the study of the genomic ‘grammar’ of a given disease using GWAS data from several human population groups may allow for the understanding of the multifunctional biological pathways that exist and their diversity observed among populations. The variation of 302 common genetic targets for endometriosis is analyzed on a set of geographically separated populations using GWAS information from the ‘1000 Genomes Project’. Using the allele frequencies, the genetic differentiation among the five population groups is shown and the common low and high frequency SNP targets are located. By associating specific SNPs clustering with low and high frequencies among the population groups selected, conclusions are drawn to the effect of population geographical translocation, epigenetic and environmental factors and regionality of the disease. It remains to be elucidated whether patients with endometriosis of different ancestral backgrounds have clear differences in clinical presentation and disease course. Thus, a more in-depth understanding of some of the molecular differences between populations, both at the genetic and protein expression levels may prove valuable, as it may suggest strategies through which treatment could be personalized. As a consequence, it may then be possible to determine the unique characteristics of the patient's disease and select the most beneficial treatment based upon a patient's individual biology.

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Authors' contributions

LP was involved in the conceptualization and methodology of the study, in obtaining resources, data curation and investigation, formal analysis, and validation as well as in the writing of the original draft, and in the reviewing and editing of the manuscript. AA was involved in data curation, investigation, validation and formal analysis, as well as in the writing of the original draft, and in the reviewing and editing of the manuscript. MZ was involved in data validation and formal analysis, and in the writing of the original draft. DV was involved in data validation, in the provision of resources and in data curation. GNG was involved in the provision of resources, in data validation, as well as in the writing of the original draft. EE was involved in the conceptualization and methodology of the study, in data curation, investigation, validation and formal analysis, in the provision of resources and project administration, in the writing of the original draft and in the reviewing and editing of the manuscript, as well as in the study supervision. LP and EE confirm the authenticity of all the raw data. All authors have read and approved the final manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References