br binding transcription factor activity br
binding, transcription factor activity,
transcriptional activity than
signal transducer activity, growth factor
signaling, positive regulation of
tissue remodeling, response to
poly-adenylate sequence within
reduced VDR level in BsmI (B)
allele versus GG (bb) genotype.36
Abbreviations: dbSNP ¼ Single Nucleotide Polymorphism Database; MAPK ¼ mitogen-activated protein kinase; OPG ¼ osteoprotegerin; PKB ¼ protein kinase B; RANKL ¼ receptor activator of nuclear factor kB ligand; TNF ¼ tumor necrosis factor; UTR ¼ untranslated region; VDR ¼ vitamin D receptor. aMajor and minor Prostaglandin J2 according to their frequencies in our control population.
We performed functional annotation analysis by conducting Go-enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses (https://www.genome.jp/kegg/) as well as protein interactions regarding our selected genes using David Bioinformatics Resources 6.8 (National Institute of Allergy and Infectious Diseases/National Institutes of Health) via the Gene ID Conversion Tool37 (https://david.ncifcrf.gov/conversion.jsp? VFROM¼NA).
We sought evidence of association of each of the 6 SNPs, and their interactions and combinations with BC risk in a multistep process. At the first stage, we calculated crude allele and genotype frequencies for each individual polymorphism and evaluated it via Hardy-Weinberg equilibrium (HWE) using a chi-square test with 1 degree of freedom among controls and cases. For testing the asso-ciation of each SNP with BC risk, a multiple logistic regression model was used for calculating odds ratios (ORs), 95% confidence
intervals (CIs), and corresponding P value of codominant, domi-nant, recessive, and log-additive models, controlling for age and family history of BC as covariates. In all analyses, the major allele or the common homozygote genotype in the control population were defined as the reference category. Akaike information criterion was also used to select the best genetic effect for each SNP.
At the second stage, we used a statistically oriented approach to examine the epistatic effects between selected SNPs, where pairwise SNP-SNP interactions were investigated using multivariate logistic models with adjustment for age and family history of BC. Because there is large number of interactions analyzed that could lead to false-positive results, Bonferroni correction of P value was con-ducted. Allelic combination (haplotype) and linkage disequilibrium (LD) analyses were also performed. Haplotypes with total frequency (pooled frequency of controls and BC cases) of > 0.05 were only considered, while rare haplotypes (total frequency < 0.05) were omitted. The overall difference in the frequency distribution of haplotypes between controls and BC cases (global test), and each haplotype association test was estimated, controlling for age and r> family history of BC as covariates. Interaction of haplotypes with pathologic data was also examined.
Concerning SNPs, all tests were carried out by SNPStats online software (Inistitut Català d’Oncologia, Barcelona, Spain; https:// www.snpstats.net/start.htm), which uses the maximum likelihood method.38 For demographic and clinical data of cases and controls, t test and the Fisher exact test were conducted for analyzing quantitative and categorical variables, respectively, using GraphPad Prism 5 (GraphPad Software, La Jolla, CA). P < .05 was considered significant.
The selected background characteristics of the study subjects are listed in Table 2. There were no significant differences between the BC and control groups with respect to menstrual history (P ¼ .15). The average age of patients in the BC group was slightly higher than in the control group, but the difference was not significant (P ¼ .057). The number of BC patients with a family history of BC was significantly higher in the BC group than in the control group (P < .0001).
Regarding the pathologic data of BC cases (Table 2), 88.7% of tumors were invasive ductal tumors. Seventy-three percent of cases were tumor, node, and metastases stages II and III, and 74.8% were small tumors (< 5 cm). Eighteen percent of cases were positive for estrogen receptor (ER)/progesterone receptor (PR) hormones. Bone metastasis was present in 28% of BC cases; all had small tumor size, and 29% of them had positive ER/PR status.
Single Locus Effects of Studied SNPs
The minor allele frequencies (MAF) of all SNPs were > 0.1. MAFs of some SNPs in controls were similar to the global MAF reported in the National Center for Biotechnology Information Single Nucleotide Polymorphism Database (dbSNP) or Trans-Omics for Precision Medicine (TOPMed), or were close to the highest population MAF reported in Ensembl GRCh37 release 89 (2017) for others (Supplemental Table 1 in the online version). The genotype distribution of OPG-rs2073618, OPG-rs2073617, and VDR FokI-rs2228570 in controls and cases; CHI3L1-rs4950928 in controls; and VDR BsmI-rs1544410 in cases followed HWE (P >