ERAP1 polymorphisms interactions and their association with Behçet's disease susceptibly: Application of Model-Based Multifactor Dimension Reduction Algorithm (MB-MDR).

BACKGROUND: Behçet's disease (BD) is a chronic multi-systemic vasculitis with a considerable prevalence in Asian countries. There are many genes associated with a higher risk of developing BD, one of which is endoplasmic reticulum aminopeptidase-1 (ERAP1). In this study, we aimed to investigate the interactions of ERAP1 single nucleotide polymorphisms (SNPs) using a novel data mining method called Model-based multifactor dimensionality reduction (MB-MDR).
METHODS: We have included 748 BD patients and 776 healthy controls. A peripheral blood sample was collected, and eleven SNPs were assessed. Furthermore, we have applied the MB-MDR method to evaluate the interactions of ERAP1 gene polymorphisms.
RESULTS: The TT genotype of rs1065407 had a synergistic effect on BD susceptibility, considering the significant main effect. In the second order of interactions, CC genotype of rs2287987 and GG genotype of rs1065407 had the most prominent synergistic effect (β = 12.74). The mentioned genotypes also had significant interactions with CC genotype of rs26653 and TT genotype of rs30187 in the third-order (β = 12.74 and β = 12.73, respectively).
CONCLUSION: To the best of our knowledge, this is the first study investigating the interaction of a particular gene's SNPs in BD patients by applying a novel data mining method. However, future studies investigating the interactions of various genes could clarify this issue.

Introduction

Behçet’s disease (BD) is a chronic vasculitis presented with multi-systemic signs and symptoms; however, it is majorly separated from other autoimmune diseases by characteristic bipolar aphthosis [1]. With a wide range of prevalence worldwide (from 0.64 per 100,000 in the UK to 420 per 100,000 in Turkey), BD is mostly distributed in countries alongside the Silk Road [2]. According to the considerable prevalence and morbidity of BD in Asian countries, understanding BD’s pathophysiology might lead to new therapeutic options and increasing patients’ quality of life. Years of research have proven that similar to many other rheumatic disorders, genetic factors have a significant role in BD’s course [3].

HLA region has been proven to have a pivotal contribution to the genetic component of BD [4]. BD’s association with HLA-B*51 is proved by several influential studies, including a meta-analysis on 4800 patients that has shown individuals with this allele have an odds ratio of 5.78 for developing BD [5]. In addition to HLA-B*51, studies have suggested a link between BD and other genes such as interleukin 10 (IL-10) and IL-23 receptor (IL-23R), some of which are associated with HLA-B*51 [6]. In our previous study, we have shown that the endoplasmic reticulum aminopeptidase-1 (ERAP1) gene polymorphisms are associated with HLA-B*51, resulting in higher BD susceptibility [7]. ERAP1 is an amino-peptidase responsible for the N-terminal trimming of peptides, which is a critical step in peptides processing and their presentation by MHC-I [8].

Furthermore, ERAP1 takes part in cleaving proinflammatory cytokine receptors such as tumor necrosis factor receptor (TNFR1) from the cell membrane [9]. Polymorphisms of ERAP1 might alternate the activity of the protein and subsequently changing the structure of peptidome available to HLA-B*51. However, the association of ERAP1 single nucleotide polymorphisms (SNPs) and BD susceptibility is not entirely clear, and some studies suggest contradictory findings, which need to assess by more comprehensive studies [7, 10, 11].

Up to now, logistic regression for high dimensional and sparse data, parameter estimation is a costly and non-accurate procedure that introduces significant standard errors because sample sizes are too small compared to the order of interaction size. Also, conventional approaches (e.g., logistic regression) used for the analysis of genomic data are oversimplified and usually cannot consider all possible associations between multiple polymorphisms and gene-gene interactions [12]. Multifactor Dimensionality Reduction (MDR) approach is now a reference in the epistasis and SNPs interactions detection field. However, MDR suffers from some significant drawbacks, including that crucial interactions could be missed owing to pooling too many cells together or that proposed MDR analysis will only reveal at most one significant epistasis model, the selection being based on computationally demanding cross-validation and permutation strategies. To overcome the aforementioned hurdles, model-based multifactor dimensionality reduction (MB-MDR) is a flexible framework to detect gene-gene or SNP-SNP interactions. MB-MDR is a non-parametric data mining method that has sufficient power and is capable of investigating the interaction of the unlimited number of genes and polymorphisms [13]. Therefore, we aimed to use the MB-MDR method to identify the interactions of ERAP1 polymorphisms and their association with BD susceptibly.

Methods

Study participants

The present study included 748 BD patients who were referred to the outpatient BD clinic in the Rheumatology Research Center, Shariati Hospital, Tehran, Iran. The International Criteria confirmed patients’ diagnosis for Behçet’s Disease (ICBD), and patients who were less than 16 years old or related to each other were excluded from the study [14, 15]. For the control group, we have included 776 healthy individuals with no clinical presentation or family history of any rheumatic disorders or autoimmune diseases, who were matched for sex, age, and ethnicity [16]. Written informed consent was obtained from all individuals themselves or their parents in cases with the age of under 18. The ethical committee of Tehran University of Medical Sciences approved the study protocol, and the relevant university guidelines did all experiments.

DNA preparation and SNP genotyping

A peripheral blood sample was collected from all participants into EDTA-anticoagulated tubes using venipuncture. Genomic DNA was extracted using the standard phenol/chloroform method, and the extracted DNA samples were stored at −20 °C. Approximately 20 ng of the genomic DNA in each sample was used for genotyping. We assessed 10 common missense SNPs from our previous study [7] that were identified in the super-population of the 1000 Genomes project and had a minor allele frequency of more than one percent (). We have also included an intronic SNP (rs1065407) that has been associated with BD in another study [17]. MGB-TaqMan Allelic Discrimination technique was used for SNP genotyping (Applied Biosystems, Foster City, CA, USA). Ten μl of reaction volumes, containing 0.25 μl of distilled water, 4.5 μl of genomic DNA, 0.25 μl of TaqMan genotyping assay mix, and 5 μl of the TaqMan genotyping master mix was used for amplification. The StepOnePlus Real-Time PCR System (Applied Biosystems) and the manufacturer’s protocol were used for genotyping the patients and healthy individuals’ samples. The allelic call was done using SDS v.1.4 software (Applied Biosystems) and the analysis of allelic discrimination plots. Finally, the genetic makeup of SNPs for each subject was considered as the genotype of that SNP.

Statistical methods

The continuous variables were indicated as mean ± SD. Allelic and genotypic frequencies of the ERAP1 SNPs were mentioned as N (%). The genotype distributions of SNPs were tested for deviation from Hardy-Weinberg equilibrium (HWE) in the control group. P-values were corrected for multiple comparisons by the Benjamini-Hochberg approach [18]. Since calculations of the main effect of ERAP1 SNPs were not available by the model-based multifactor dimensionality reduction (MB-MDR), multiple logistic regression has been used to obtain the main effects of ERAP1 SNPs, simultaneity. To adjust for main effects, main effects should be calculated. MB-MDR has been proposed by Calle et al. as a dimension reduction method for exploring SNP-SNP interactions with disease susceptibly in case-control association studies [19]. MB-MDR method has proven to be more potent than multifactor dimensionality reduction (MDR) in the presence of genetic heterogeneity [20]. MB-MDR can unify the best of both nonparametric and parametric machine learning algorithms.

On the other hand, characterization, and identification SNP-SNP interactions lack performance in the absence of proper statistical methods and large sample sizes. Logistic regression, as a standard tool for modeling effects and interactions with binary response data, lacks power in the identification of gene interactions in high-order levels due to sparsity and separation [21]. Thus, in this study, SNP-SNP interactions were calculated by the MB-MDR algorithm. MB-MDR shows high power in the presence of all types of noises, such as missing data, genotyping error, genetic heterogeneity, and low sample size [22]. This algorithm was performed by “mbmdr” R package version 3.5.1. To assess the significance in MB-MDR, permutation test with 1000 replications has been done, which corrects for multiple testing (overall marker pairs) and adequately controls the family-wise error rate at α = 0.05.

Results

In this case-control study, 748 patients and 776 age-, sex-, and ethnicity- matched healthy controls were included according to the inclusion and exclusion criteria [16]. In BD patients, the mean age was 40.26 ± 10.88 years, and in the control group was 38.88 ± 11.54 years (P-value = 0.076). Out of 748 patients and 776 healthy individuals, 448 (59.9%) and 476 (61.3%) were male, respectively (P-value = 0.599). Based on the results of assessing the main effects of ERAP1 SNPs, the TT genotype of rs1065407 SNP (β = 0. 23, and adjusted P-value = 0.034) had a significant synergistic effect on BD. The synergistic effect of an allele is described as the allele increasing the disease risk, and the antagonistic effect is described as the allele having a protective effect regarding the disease susceptibility. In contrast, TT genotype of rs30187 SNP (β = -0.26 and adjusted P-value = 0.041) and AA genotype of rs469876 SNP (β = -0.20 and adjusted P-value = 0.046) had significant antagonistic effects on BD (). Other ERAP1 SNPs do not have significant main effects concerning BD susceptibly.

Table 2 summarizes the results of SNP-SNP interactions for six important SNPs (rs1065407, rs30187, rs469876, rs2287987, rs17482078, and rs26653). Based on the results of second-order interaction effects, there were only six significant 2-locus models. For instance, CC genotype of rs2287987 and GG genotype of rs1065407 (β = 12.74 and adjusted P-value = 2.12×10−10) had a significant synergistic effect on BD susceptibility. rs30187 and rs1065407, CT, and TT genotype (β = -0.39 and adjusted P-value of 1.98×10−3) had a significant antagonistic effect on BD. Synergistic effects of rs469876 (AA and GG) genotypes with rs1065407 (GG and GT) genotypes were significant as well (β = 0.32, adjusted P-value = 4.73×10−3). Effects of rs30187 and rs469876 (CC vs. AA) and (TT vs. AG) were also significantly synergistic (β = 0.32 adjusted P-value = 2.39×10−2). rs26653 (CC) with rs1065407 (GG) had a significant synergistic effect on BD (β = 0.76, adjusted P-value = 2.49×10−2). However, the results of rs26653 (CT) and rs469876 (AG) showed a significant negative association with BD susceptibly (β = -0.42, adjusted P-value = 7.38×10−2).

Table 2

Model-based multifactor dimensionality reduction algorithm for assessing the main and interaction effects of 11 ERAP1 SNPs on Behçet’s disease risk (748 Iranian BD patients and776 healthy individuals).

Order	Significant Effects	Synergistic Effect				Antagonism Effect				Permutation Test
		N. levels	Genotypes	Coefficient	Adj. P-value	N. levels	Genotypes	Coefficient	Adj. P-value	Perm. P-value
Main Effects	rs1065407	1	TT	0.23	0.034	0	NA	NA	NA	0.019
	rs30187	0	NA	NA	NA	1	TT	-0.26	0.041	0.18
	rs469876	0	NA	NA	NA	1	AA	-0.20	0.046	0.054
#2 order interactions	rs2287987+rs1065407	1	CC+GG	12.74	2.12×10−10	0	NA	NA	NA	0.065
	rs30187+rs1065407	0	NA	NA	NA	1	CT+TT	-0.39	1.98×10−3	0.053
	rs469876+rs1065407	2	AA+GG GG+GT	0.32	4.73×10−3	0	NA	NA	NA	0.181
	rs30187+rs469876	2	CC+AA TT+AG	0.32	2.39×10−2	0	NA	NA	NA	0.091
	rs26653+rs1065407	1	CC+GG	0.76	2.49×10−2	0	NA	NA	NA	0.210
	rs26653+rs469876	2	CC+AA GG+AG	0.54	2.83×10−2	1	CT+AG	-0.42	7.38×10−2	0.193
#3 order Interaction	rs1065407+rs2287987+rs26653	1	GG+CC+ CC	12.74	2.13×10−10	0	NA	NA	NA	0.243
	rs1065407+rs2287987+rs30187	1	GG+CC+ TT	12.73	2.15×10−10	1	TT+ CT+ CT	-0.39	5.95×10−2	0.230
	rs1065407+rs30187+rs469876	3	GG+TT+AG GT+CT+AA	0.43	2.87×10−2	1	TT+ CT+ AG	-0.67	1.26×10−3	0.169
	rs30187+ rs1065407+rs26653	4	CC+GG+CC TT+GT+GG	0.77	2.36×10−2	0	NA	NA	NA	0.137
	rs1065407+rs2287987+rs469876	2	GG+CC+GG GT+TT+AA	0.04	9.77×10−1	1	TT+CT+AG	-0.92	3.18×10−2	0.229
#4 order Interaction	rs1065407+rs2287987+rs30187+rs26653	7	GG+CC+ CC+CC GT+CT+ TT+CG	0.53	1.94×10−1	2	TT+TT+CT+ GG GT+CT+TT+CG	-0.88	7.50×10−3	0.184
	rs1065407+rs2287987+rs26653+rs469876	5	GG+CC+ CC+GG GT+CT+ GG+AA	0.66	4.49×10−1	2	TT+TT+ CG+AG GT+CT+GG+AG	-0.65	1.18×10−2	0.219
#5 order Interaction	rs1065407+rs2287987+rs30187+rs26653+rs17482078	11	GT+TT+ CC+CC+ TT TT+CT+TT+CG+ CT	0.32	3.93×10−1	2	TT+CT+ CT+GG+ CT GT+TT+ TT+CG+ TT	-0.89	7.25×10−3	0.032

Considering third-order interaction effects, we had five 3-locus models for SNP-SNP interactions of ERAP1 SNPs. For example, the GG genotype of rs1065407, CC genotype of rs2287987, and CC genotype of rs26653 had a significant synergistic effect on BD by a 3-locus model (β = 12.74, adjusted P-value = 2.13×10−10). However, the 3-locus model (rs1065407, rs2287987, rs26653) did not have any significant antagonistic effect on BD. Considering rs1065407, rs2287987, and rs30187, results reveal that the synergistic effect of (GG, CC, and TT) genotypes and the antagonistic effect of (TT, CT and CT) genotypes on BD, were significant as well. Besides, rs1065407 (TT), rs30187 (CT) and rs469876 (AG) had a significant antagonistic effect on BD (β = -0.67, adjusted P-value = 1.26×10−3). In addition, rs1065407 (TT), rs2287987 (CT) and rs469876 (AG) interaction had a significant antagonistic effect on BD (β = -0.92, adjusted P-value = 3.18×10−2). In contrast, (rs1065407: GG, rs30187: TT, rs469876: AG), (rs1065407: GG, rs2287987: CC, rs469876: GG), and (rs30187: CC, rs1065407: GG, rs26653: CC) had significant synergistic effects on BD. More details are shown in the third-order interaction section of Table 2.

Results of fourth-order interaction effects indicated that (rs1065407: GG, rs2287987: CC, rs30187: CC, rs26653: CC) and (rs1065407: GG, rs2287987: CC, rs26653: CC, rs469876: GG) had significant synergistic effects on BD. In contrast, (rs1065407: TT, rs2287987: TT, rs30187: CT, rs26653: GG) and (rs1065407: TT, rs2287987: TT, rs26653: CG, rs469876: AG) had significant antagonistic effects on BD. Based on the results of five-order interaction effects, (rs1065407: GT, rs2287987: TT, rs30187: CC, rs26653: CC, rs17482078: TT) had a significant synergistic effect on BD (β = 0.32, adjusted P-value = 3.93×10−1). However, (rs1065407: TT, rs2287987: CT, rs30187: CT, rs26653: GG, rs17482078: CT) had a significant antagonistic effect on BD (β = -0.89, adjusted P-value = 7.25×10−3). In six- order interaction effects, no significant effects were observed ().

More details of the results of 11 ERAP1 SNP-SNP interactions are presented in the supplementary Table. Also, the entropy-based interaction network of 11 ERAP1 SNPs was shown in by using MDR. To assess the sensitivity and cross-validity of the results of MB-MDR, permutation results are shown in the last column of Table 2.

Discussion

In this study, we aimed to investigate the interactions of the ERAP1 gene polymorphisms and their associations with BD susceptibility in an Iranian cohort. Using the MB-MDR package, we have found plenty of synergistic and antagonistic significant interactions between ERAP1 polymorphisms and BD development. Considering the main effects, the TT genotype of rs1065407 had a synergistic effect on BD susceptibility. In the second-order interactions, CC genotype of rs2287987 and GG genotype of rs1065407 had the most prominent synergistic effect (β = 12.74). Furthermore, the mentioned genotypes also had significant interactions with CC genotype of rs26653 and TT genotype of rs30187 in the third-order (β = 12.74 and β = 12.73, respectively). Hence, we propose that the genotypes, as mentioned earlier of rs2287987, rs1065407, rs26653, and rs30187, could have prominent interactions resulting in a higher risk of developing BD.

ERAP1 gene is located in the 5q15 chromosome, and its expression has been observed in many tissues [23]. There are two main processes that ERAP1 is proposed to have a role in them. First, this amino-peptidase is involved in optimizing the length of peptides to bind with MHC-class I molecules by trimming their N-terminal in the endoplasmic reticulum (ER) [23]. Moreover, ERAP1 is involved in the cleavage process of various cytokine receptors such as TNFR1, Interleukin 1 receptor II (IL-1RII), and Interleukin 6 receptor α (IL-6 α), which results in receptor shedding [24, 25]. Previous studies have shown that the ERAP1 gene is associated with other autoimmune disorders such as ankylosing spondylitis (AS) and psoriasis [26, 27]. Homozygosity of ERAP1 polymorphisms is proposed to be correlated with a lower risk of AS and psoriasis, whereas it might be associated with a higher risk of developing BD [28, 29]. These differences could be justified by the fact that loading different peptides on MHC-class I molecules can alter the subsequent immune response.

Our results indicated that the homozygous genotypes of minor alleles of rs2287987, rs1065407, rs26653, and rs30187 had the most prominent interactions causing BD susceptibility. In this regard, it has been demonstrated that the frequencies of the homozygous alleles of the ERAP1 gene are higher among BD patients [11]. As it was shown in further studies, these combinations of homozygote ERAP1 SNPs could result in alternations in the surface electrostatic potential of the protein [30]. These changes might alter the trimming activity of ERAP1, resulting in an altered composition of peptidome that is available for binding to HLA-B*51. This claim could support the higher risk of developing BD observed in individuals carrying the mentioned genotypes. Furthermore, some SNPs such as rs30187 (Arg528Lys) are placed proximal to the entrance pocket of the protein [28]. Amino acid changes in such positions could modify the ideal structure of the protein and alter the enzyme activity.

Although several studies have investigated the association of ERAP1 polymorphisms and BD, there have been some contradictory findings that motivated us to utilize a more complex statistical method for addressing this issue. Zhang et al. evaluated 930 Chinese patients and proposed that rs1065407 and rs10050860 might be associated with increased risk of BD [17]. Sousa and colleagues studied another Iranian cohort and proposed that rs10050860 and rs13154629 might contribute to the genetic susceptibility of BD [15]. Moreover, Conde-Jaldón et al. found that homozygous genotypes for the minor alleles of rs27044, rs10050860, rs30187, and rs2287987 could be considered as risk factors for BD [10]. Takeuchi and colleagues found a haplotype consisting of 10 SNPs (five of which were non-ancestral), which was associated with a higher risk of developing BD, especially in those individuals who carry HLA-B*51 [30]. Interestingly, our results indicated that homozygote genotypes of minor alleles of rs30187 and rs2287987 are associated with a higher risk of BD. rs30187 and rs2287987 are among those five SNPs that their non-ancestral alleles were mentioned in Takeuchi’s study. Finally, the previous study by our team and the study on the Turkish population revealed that ERAP1 polymorphisms have epistatic interactions with HLA-B*51 contributing to BD risk [7, 30].

In conclusion, this is the first study investigating the interaction of a particular gene’s SNPs in BD patients by applying a novel data mining method (MB-MDR package). Model-Based MDR as a flexible framework and a reference method to detect gene–gene or SNP-SNP interactions has adequate power even the presence of genotyping errors, missing genotypes, and genetic heterogeneity in this study compare with traditional methods (e.g., logistics regression). Finally, a significant interaction between minor genotypes of ERAP1 polymorphisms was observed in BD patients in comparison to healthy individuals. rs2287987, rs1065407, rs26653, and rs30187 interactions had the strongest association with developing BD in our study population. Taken together, these findings imply the contribution of ERAP1 polymorphisms in BD pathogenesis. However, further studies investigating the interactions of different genes could shed more light on this issue.

(DOCX)

Click here for additional data file.

(RAR)

Click here for additional data file.

12 Nov 2019

PONE-D-19-28263

ERAP1 polymorphisms interactions and their association with Behçet’s disease susceptibly: Application of Model Based Multifactor Dimension Reduction Algorithm (MB-MDR)

PLOS ONE

Dear Dr Mahmoudi,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Dec 27 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Zezhi Li, Ph.D., M.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

1. In your Data availability statement, you wrote, 'Data are available upon request.' Please include the contact information for data requests.

2. Thank you for including your funding statement; "NO. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

Please provide an amended Funding Statement that declares *all* the funding or sources of support received during this specific study (whether external or internal to your organization) as detailed online in our guide for authors at http://journals.plos.org/plosone/s/submit-now.

Please state what role the funders took in the study.  If any authors received a salary from any of your funders, please state which authors and which funder. If the funders had no role, please state: "The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript."

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

3. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Behçet’s disease (BD) is a chronic multi-systemic vasculitis with a considerable prevalence in the Asian countries. Riahi et al ’s manuscript investigate interactions of ERAP1 single nucleotide polymorphisms (SNPs) using a novel data mining method called Modelbased multifactor dimensionality reduction (MB-MDR). They have included 748 BD patients and 776 healthy controls. Their results indicated that TT genotype of rs1065407 had a synergistic effect on BD susceptibility, considering the significant main effect. In the second order of interactions, CC genotype of rs2287987 and GG genotype of rs1065407 had the most prominent synergistic effect (β=12.74). The mentioned genotypes also had significant interactions with CC genotype of rs26653 and TT genotype of rs30187 in the third order (β=12.74 and β=12.73, respectively). In general, their results support their statement. However, there are still some issue need to be resolved.

1. It will be easy to understand the manuscript if the author explain how they define genotype.

2. It will be interesting if the author explain difference between MB-MDR and other analysis methods?

3. I am wondering what is the advantage of MB-MDR compared with other methods when they analyze data?

4. I am wondering if the author will get similar results(interactions of ERAP1 SNPs) by using other analysis method?

Reviewer #2: Comments for “ERAP1 polymorphisms interactions and their association with Behçet’s disease susceptibly: Application of Model Based Multifactor Dimension Reduction Algorithm (MB-MDR)”

In this study, Parisa, et al. evaluated the potential synergistic and antagonism effect of ERAP1 SNPs on patients with Behçet’s disease (BD) by using a new method MB-MDR. The analysis results are based on a considerable number of cases (748) and healthy controls (776), which could comprehensively assess the correlation between ERAP1 SNPs and disease occurrence in the area surveyed, thus, it supplies a way to predict the risk of disease. While the mechanism of Behçet’s disease is uncertain and multiple factors could contribute to this disease, animal models for this disease is unavailable at present, acquisition of the statistics from a clinical sample is the only way we approach this disease. Therefore, our analysis methods largely determine the reliability of statistic. This study provides the data calculated by the “mbmdr” R package version 3.5.1, the result could be insufficiently supported due to the following three major concerns.

1. Although this method MB-MDR is powerful, it doesn’t means its results are more reliable than other methods. Because the data are confront with other research data, for example, conclusions from Kirino, et al. 2013 and your former conclusion in Mahmoudi, et al. 2018.

2. Since DB is a chronic disease caused by many factors, evaluation of the synergism of SNPs in individual gene could not meaningful, while the synergism of SNPs between or among multiple genes could be more reliable.

3. In this study, there are only tables, which are not direct-viewing diagrams, especially for the Table 2, if authors draw a diagram showing the main and interaction effects, it could be better for understanding by the readers.

Reviewer #3: 1. In the Introduction section, I would like to the author of this paper to give a more through description of how ERAP1 gene polymorphisms are associated with HLA-B*51.

2. In the Discussion section, the author should discuss the advantage of MB-MDR method

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

15 Dec 2019

Manuscript ID: PONE-D-19-28263 entitled “ERAP1 polymorphisms interactions and their association with Behçet’s disease susceptibly: Application of Model-Based Multifactor Dimension Reduction Algorithm (MB-MDR)”

Dear academic editor,

Zezhi Li,

We are grateful to the reviewers and the editorial board of PLOS ONE, for their constructive criticisms on our paper. We revised the manuscript accordingly. The modifications are given as point-by-point responses to the comments of the reviewers. All changes in the manuscript are highlighted in yellow color. Also, all authors declare that they are agreed on the revision. We wish to thank the comments and hope that the revised version of the manuscript may be now suitable for publication.

Yours sincerely,

Anoshirvan Kazemnejad: Professor of Biostatistics, Department of Biostatistics, Faculty of Medical Sciences, Tarbiat Modares University, P.O.Box: 14115-111, Tehran, IR, Iran. E-mail address: kazem_an@modares.ac.ir

Mahdi Mahmoudi: Rheumatology Research Center, Tehran University of Medical Sciences, P.O. Box: 14117-13137, Tehran, IR. E-mail address: mahmoudim@tums.ac.ir

Comments to the Author

Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Behçet’s disease (BD) is a chronic multi-systemic vasculitis with a considerable prevalence in Asian countries. Riahi et al.’s manuscript investigates interactions of ERAP1 single nucleotide polymorphisms (SNPs) using a novel data mining method called Model-based multifactor dimensionality reduction (MB-MDR). They have included 748 BD patients and 776 healthy controls. Their results indicated that the TT genotype of rs1065407 had a synergistic effect on BD susceptibility, considering the significant main effect. In the second order of interactions, CC genotype of rs2287987 and GG genotype of rs1065407 had the most prominent synergistic effect (β=12.74). The mentioned genotypes also had significant interactions with CC genotype of rs26653 and TT genotype of rs30187 in the third-order (β=12.74 and β=12.73, respectively). In general, their results support their statement. However, there is still some issue that needs to be resolved.

1. It will be easy to understand the manuscript if the author explains how they define genotype.

Response: Thanks for your comment. We have added the explanation of genotypes to the method section.

2. It will be interesting if the author explains the difference between MB-MDR and other analysis methods?

Response: As mentioned at the end of the introduction section, conventional approaches (e.g., stepwise logistic regression) used for the analysis of genomic data are oversimplified and usually cannot consider all possible significant interactions between multiple polymorphisms. For example, logistics regression for identifying the most important of interactions between multiple polymorphisms is not an appropriate method in a large number of dichotomous variables (or SNPs), small sample size, and sparse data (1, 2). In logistic regression for high dimensional and sparse data, parameter estimation is a costly and non-accurate procedure that introduces large standard errors because sample sizes are too small compared to the order of interaction size. As a consequence, many false positives are generated when dealing with such data (3). In other words, logistic regression performs poorly when there is a dimensionality problem (4). With no ‘best’ statistical approach available, full combinatorial approaches (e.g., Multifactor Dimensionality Reduction) may be optimal for detecting SNPs interactions. MDR approach is now a reference in the epistasis and SNPs interactions detection field. No parameters are estimated (i.e., nonparametric), and no assumptions are made on the genetic model (i.e., model-free) under this supervised classification approach. However, MDR suffers from some major drawbacks, including that crucial interactions could be missed owing to pooling too many cells together or that proposed MDR analyses will only reveal at most one significant epistasis model, the selection being based on computationally demanding cross-validation and permutation strategies (5). To overcome the aforementioned hurdles, Model-Based MDR is a flexible framework to detect gene–gene or SNP-SNP interactions. Besides, MB-MDR has adequate power, even the presence of error sources or noise/ genotyping errors, missing genotypes, phenotypic mixtures, and genetic heterogeneity (5). The mentioned details were added in the introduction section.

3. I am wondering what is the advantage of MB-MDR compared with other methods when they analyze data?

Response: As mentioned in the response of the previous comment, Model-Based MDR is a flexible framework to detect gene–gene or SNP-SNP interactions with adequate power even the presence of error sources or noise/ genotyping errors, missing genotypes, phenotypic mixtures and genetic heterogeneity (5). Please see the before comment’s answer.

4. I am wondering if the author will get similar results (interactions of ERAP1 SNPs) by using other analysis methods?

Response: Based on the previous comment’s answer, logistic regression has a high computational burden and non-accurate estimations with large standard errors. However, the results of MDR were reported as follow:

Figure: Entropy-based network of interactions between12 ERAP1 SNPs.

For example, the results of entropy in MDR:

Attribute H(A) H(A|C) I(A; C)

--------- ---- ------ ------

Y 1 1 0

rs1065407 1.4731 1.4196 0.0534

rs2287987 0.9805 0.8194 0.1611

rs30187 1.518 1.4962 0.0218

rs10050860 0.8635 0.6771 0.1864

rs27044 1.2802 1.2662 0.0139

rs26653 1.486 1.4505 0.0354

rs27434 1.425 1.4107 0.0142

rs469876 1.1953 1.1863 0.009

rs17481856 0.7089 0.7075 0.0014

rs28096 1.4858 1.4455 0.0402

rs13167972 1.5345 1.5059 0.0285

Attribute A Attribute B H(AB) H(AB|C) I(A;B) IG(A;B;C) I(AB;C)

----------- ----------- ----- ------- ------ -------- -------

rs1065407 Y 0.9889 0.9489 1.4842 -0.0136 0.0399

rs2287987 Y 0.7689 0.606 1.2116 0.0018 0.1629

rs2287987 rs1065407 0.7778 0.61 1.6757 -0.0467 0.1678

rs30187 Y 0.9376 0.9314 1.5804 -0.0157 0.0062

rs30187 rs1065407 1 0.9377 1.9911 -0.0131 0.0622

rs30187 rs2287987 0.7807 0.6112 1.7178 -0.0134 0.1695

rs10050860 Y 0.7659 0.5795 1.0976 0 0.1864

rs10050860 rs1065407 0.7659 0.5795 1.5706 -0.0534 0.1864

rs10050860 rs2287987 0.7748 0.6087 1.0691 -0.1813 0.1662

rs10050860 rs30187 0.7659 0.5795 1.6156 -0.0218 0.1864

rs27044 Y 0.9936 0.9899 1.2865 -0.0103 0.0037

rs27044 rs1065407 0.9906 0.9463 1.7626 -0.0231 0.0443

rs27044 rs2287987 0.7893 0.6335 1.4713 -0.0192 0.1558

rs27044 rs30187 0.9361 0.9293 1.8621 -0.029 0.0068

rs27044 rs10050860 0.7659 0.5795 1.3777 -0.0139 0.1864

rs26653 Y 0.9103 0.8887 1.5757 -0.0139 0.0216

rs26653 rs1065407 1 0.9277 1.959 -0.0165 0.0723

rs26653 rs2287987 0.7893 0.6335 1.6771 -0.0407 0.1558

rs26653 rs30187 0.9856 0.9697 2.0184 -0.0414 0.0159

rs26653 rs10050860 0.7659 0.5795 1.5835 -0.0354 0.1864

rs26653 rs27044 0.9863 0.962 1.7799 -0.0251 0.0243

rs27434 Y 0.9986 0.9914 1.4263 -0.0071 0.0072

rs27434 rs1065407 0.9996 0.9406 1.8984 -0.0087 0.059

rs27434 rs2287987 0.7921 0.643 1.6133 -0.0262 0.1491

rs27434 rs30187 0.9974 0.9891 1.9456 -0.0278 0.0083

rs27434 rs10050860 0.7659 0.5795 1.5225 -0.0142 0.1864

rs27434 rs27044 0.9974 0.9891 1.7078 -0.0199 0.0083

rs27434 rs26653 0.9961 0.9849 1.9149 -0.0385 0.0112

rs469876 Y 0.9551 0.9505 1.2402 -0.0045 0.0045

rs469876 rs1065407 0.9964 0.9549 1.6719 -0.0209 0.0415

rs469876 rs2287987 0.7807 0.6297 1.395 -0.0191 0.151

rs469876 rs30187 0.9753 0.9324 1.738 0.012 0.0429

rs469876 rs10050860 0.7659 0.5795 1.2929 -0.009 0.1864

rs469876 rs27044 0.8712 0.8636 1.6042 -0.0153 0.0076

rs469876 rs26653 0.9796 0.9206 1.7017 0.0145 0.059

rs469876 rs27434 1 0.9813 1.6203 -0.0046 0.0186

rs17481856 Y 0.9812 0.9812 0.7277 -0.0014 0

rs17481856 rs1065407 0.9889 0.9489 1.1931 -0.0149 0.0399

rs17481856 rs2287987 0.7807 0.6379 0.9087 -0.0196 0.1428

rs17481856 rs30187 0.9974 0.9901 1.2296 -0.0159 0.0073

rs17481856 rs10050860 0.7659 0.5795 0.8065 -0.0014 0.1864

rs17481856 rs27044 0.9936 0.9899 0.9955 -0.0116 0.0037

rs17481856 rs26653 0.8441 0.8053 1.3509 0.002 0.0388

rs17481856 rs27434 0.9977 0.99 1.1362 -0.008 0.0077

rs17481856 rs469876 0.9999 0.9919 0.9043 -0.0023 0.0081

rs28096 Y 0.9643 0.9297 1.5214 -0.0056 0.0346

rs28096 rs1065407 0.9993 0.9376 1.9595 -0.0319 0.0617

rs28096 rs2287987 0.7807 0.6209 1.6855 -0.0415 0.1598

rs28096 rs30187 0.9725 0.929 2.0313 -0.0186 0.0435

rs28096 rs10050860 0.7659 0.5795 1.5833 -0.0402 0.1864

rs28096 rs27044 0.9665 0.9292 1.7995 -0.0168 0.0373

rs28096 rs26653 0.9949 0.9359 1.9768 -0.0166 0.059

rs28096 rs27434 0.9706 0.9324 1.9402 -0.0163 0.0382

rs28096 rs469876 0.9842 0.9494 1.6968 -0.0144 0.0348

rs28096 rs17481856 0.9643 0.9297 1.2304 -0.007 0.0346

rs13167972 Y 0.8817 0.8593 1.6528 -0.0062 0.0224

rs13167972 rs1065407 0.9906 0.9487 2.0169 -0.0401 0.0418

rs13167972 rs2287987 0.7865 0.6323 1.7285 -0.0354 0.1542

rs13167972 rs30187 0.9771 0.9337 2.0754 -0.007 0.0434

rs13167972 rs10050860 0.7659 0.5795 1.632 -0.0285 0.1864

rs13167972 rs27044 0.8993 0.8746 1.9153 -0.0177 0.0247

rs13167972 rs26653 0.9155 0.8517 2.105 -0.0002 0.0638

rs13167972 rs27434 0.9779 0.9376 1.9815 -0.0025 0.0403

rs13167972 rs469876 0.9138 0.8916 1.816 -0.0154 0.0221

rs13167972 rs17481856 0.8712 0.8496 1.3722 -0.0083 0.0216

rs13167972 rs28096 0.9945 0.9455 2.0257 -0.0197 0.049

Reviewer #2: Comments for “ERAP1 polymorphisms interactions and their association with Behçet’s disease susceptibly: Application of Model-Based Multifactor Dimension Reduction Algorithm (MB-MDR)”

In this study, Parisa et al. evaluated the potential synergistic and antagonism effect of ERAP1 SNPs on patients with Behçet’s disease (BD) by using a new method MB-MDR. The analysis results are based on a considerable number of cases (748) and healthy controls (776), which could comprehensively assess the correlation between ERAP1 SNPs and disease occurrence in the area surveyed; thus, it supplies a way to predict the risk of disease. While the mechanism of Behçet’s disease is uncertain and multiple factors could contribute to this disease, animal models for this disease is unavailable at present, acquisition of the statistics from a clinical sample is the only way we approach this disease. Therefore, our analysis methods largely determine the reliability of the statistic. This study provides the data calculated by the “mbmdr” R package version 3.5.1, the result could be insufficiently supported due to the following three major concerns.

Response: Thanks for your comments. The mentioned corrections were done.

1. Although this method MB-MDR is powerful, it doesn’t mean its results are more reliable than other methods. Because the data are confronted with other research data, for example, conclusions from Kirino et al. 2013 and your former conclusion in Mahmoudi et al. 2018.

Response: In general, MB-MDR has different results compared to other algorithms in the mentioned studies because of the different model’s assumptions and methodology. However, typical approaches perform poorly when there is a dimensionality problem for identifying interactions in genetics studies. On the other hand, MDR reduces dimensions by converting a high-dimensional model to a one-dimensional one (4). Nevertheless, MDR as a non-parametric algorithm suffers from some major drawbacks, including that critical interactions could be missed owing to pooling too many cells together or that proposed MDR analyses will only reveal the most significant epistasis model based on computationally demanding cross-validation and permutation strategies (5). Parametric algorithm (MB-MDR) tend to have particular model assumptions, which lead to our ability to determine statistical significance under such assumptions. In a non-parametric test (MDR), we often have fewer assumptions to evaluate, but also differences in how statistical significance is determined. For example, if we run a non-parametric test, such as MDR, there is no P-value table. Based on the permutation test in cross-validation of both MDR, MB-MDR, both algorithms considered as appropriate and reference methods in the epistasis and SNPs interactions detection field (5).

2. Since DB is a chronic disease caused by many factors, evaluation of the synergism of SNPs in an individual gene could not be meaningful, while the synergism of SNPs between or among multiple genes could be more reliable.

Response: Thank you for your valuable comment. As you have mentioned, there are many factors contributing to BD’s pathophysiology, including various genes and also multiple environmental variables. Indeed, evaluating these factors, especially the gene-gene interactions, would shed more light on the underlying mechanisms of BD. However, this article is the starting point where we explored the application of a robust statistical method to investigate the interactions between ERAP1 SNPs in BD. Undoubtedly, our team is determined to test such powerful methods to evaluate the synergism of different genes’ SNPs in BD and other multifactor rheumatic diseases in future studies.

3. In this study, there are only tables, which are not direct-viewing diagrams, especially for the Table 2, if authors draw a diagram showing the main and interaction effects, it could be better for understanding by the readers.

Response: Thank you for the comment, the network based on the results of Table 2 and 3 were shown and added as Figure 1 in the results section.

Figure 1: SNP-SNP entropy-based interaction network of 12 ERAP1 SNPs.

Reviewer #3:

1. In the Introduction section, I would like the author of this paper to give a more thorough description of how ERAP1 gene polymorphisms are associated with HLA-B*51.

Response: Thank you for your comment. We have provided more explanation regarding the possible association of ERAP1 gene polymorphisms and HLA-B*51 in the introduction section.

2. In the Discussion section, the author should discuss the advantage of the MB-MDR method

Response: Thanks. The advantages of the MB-MDR method were added at the end of the discussion section.

1. Briggs F, Ramsay P, Madden E, Norris J, Holers V, Mikuls TR, et al. Supervised machine learning and logistic regression identifies novel epistatic risk factors with PTPN22 for rheumatoid arthritis. Genes and immunity. 2010;11(3):199.

2. Johnstone IM, Titterington DM. Statistical challenges of high-dimensional data. The Royal Society Publishing; 2009.

3. Niel C, Sinoquet C, Dina C, Rocheleau G. A survey about methods dedicated to epistasis detection. Frontiers in genetics. 2015;6:285.

4. Kim H, Jeong H-B, Jung H-Y, Park T, Park M. Multivariate Cluster-Based Multifactor Dimensionality Reduction to Identify Genetic Interactions for Multiple Quantitative Phenotypes. BioMed research international. 2019;2019.

5. John JMM, Van Lishout F, Van Steen K. Model-Based Multifactor Dimensionality Reduction to detect epistasis for quantitative traits in the presence of error-free and noisy data. European Journal of Human Genetics. 2011;19(6):696.

Submitted filename: Response to reviewers.doc

Click here for additional data file.

6 Jan 2020

ERAP1 polymorphisms interactions and their association with Behçet’s disease susceptibly: Application of Model-Based Multifactor Dimension Reduction Algorithm (MB-MDR)

PONE-D-19-28263R1

Dear Dr. Mahmoudi,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Zezhi Li, Ph.D., M.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Behçet’s disease (BD) is a chronic multi-systemic vasculitis with a considerable prevalence in the Asian countries. Riahi et al ’s manuscript investigate interactions of ERAP1 single nucleotide polymorphisms (SNPs) using a novel data mining method called Modelbased multifactor dimensionality reduction (MB-MDR). They have included 748 BD patients and 776 healthy controls. Their results indicated that TT genotype of rs1065407 had a synergistic effect on BD susceptibility, considering the significant main effect. In the second order of interactions, CC genotype of rs2287987 and GG genotype of rs1065407 had the most prominent synergistic effect (β=12.74). The mentioned genotypes also had significant interactions with CC genotype of rs26653 and TT genotype of rs30187 in the third order (β=12.74 and β=12.73, respectively). In general, they answer my questions.

Reviewer #2: (No Response)

Reviewer #3: The authors of this study did a thorough revision work. The authors have addressed all the issues I was concerned during the initial submission. I am satisfied with their revision work.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

24 Jan 2020

PONE-D-19-28263R1

ERAP1 polymorphisms interactions and their association with Behçet’s disease susceptibly: Application of Model-Based Multifactor Dimension Reduction Algorithm (MB-MDR)

Dear Dr. Mahmoudi:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Zezhi Li

Academic Editor

PLOS ONE

Table 1

Allele frequencies of 11 ERAP1 SNPs.

SNP	Position on chromosome five	Alleles	Amino acid changes	Minor allele frequency, %		P value	Odd ratio (95% confidence interval)
				cases	controls
rs1065407	96,776,379	T > G	Intronic	36.6	32.5	0.018	1.20 (1.03–1.39)
rs27044	96,783,148	C > G	Glu730Gln	28.5	29.1	0.74	0.97 (0.83–1.14)
rs17482078	96,783,162	C > T	Arg725Gln	12.6	10.3	0.052	1.25 (1.00–1.56)
rs10050860	96,786,506	C > T	Asp575Asn	12.5	10.1	0.039	1.27 (1.01–1.59)
rs30187	96,788,627	C > T	Arg528Lys	40.1	39.7	0.82	1.02 (0.88–1.18)
rs2287987	96,793,832	T > C	Met349Val	12.5	10.2	0.040	1.27 (1.01–1.59)
rs27895	96,793,840	C > T	Gly346Asp	9.8	9.9	0.98	1.00 (0.79–1.26)
rs26618	96,795,133	T > C	Ile276Met	20.1	22.9	0.059	0.85 (0.71–1.01)
rs26653	96,803,547	G > C	Pro127Arg	40.2	39.7	0.75	1.02 (0.89–1.18)
rs3734016	96,803,761	C > T	Glu56Lys	1.9	2.4	0.40	0.81 (0.50–1.32)
rs72773968	96,803,892	G > A	Thr12Ile	9.8	9.9	0.88	0.98 (0.77–1.25)