Based on the tetra-ARMS PCR design in this experiment, the length of DNA fragment amplified by external primers is 747 bp, the fragment amplified for C nucleotide by internal primers is 173 bp, and for T nucleotide is 611 bp (Fig. 1). If the sample studied is heterozygote for this SNP, it will show all these bands.

DNA amplification products in tetra-ARMS PCR

Based on 100 individuals (200 alleles) studied, the allele frequency of the SNP for T alleles was 0.66, while the same value for C allele was 0.36. These values were 0.62 and 0.38 for azoospermia samples, respectively. Moreover, they were 0.71 and 0.29 in control samples, respectively. These results reveal that allele frequencies obtained in Iranian sample studied differ from what has been reported from the other countries because they reported the allele C as the frequent allele and it was considered an ancestral allele (about 0.999 in Ensemble site data).

The genotype frequencies obtained showed that 67% of the samples were heterozygote (T/C), while 33% were homozygotes (TT). We obtained no CC homozygote in the studied individuals. These values for azoospermia samples were 76% T/C and 24% TT, while they were 58% and 42% in control samples, respectively. SNP-STAT analysis of data showed a significant association (p = 0.04) with the studied SNP and azoospermia in Iranian samples.

In total, we sequenced twenty samples. Accordingly, we obtained 698 bp length DNA after alignment and curation, which showed 11 polymorphic sites among individuals (Fig. 2). The Kimura 2-parameter genetic distance varied from 0.001 to 0.004 among the studied samples. The mean nucleotide diversity obtained was 0.004, and Tajima’s D statistic was D = − 0.3486. All these results indicated a low degree of nucleotide substitution and that these substitutions are not under selective pressure. The studied samples differed in sequences as they formed different clusters/clades in TCS network (Fig. 3). Some of the sequences were placed close to each other and formed three main haplotype groups (Group 1–3, in Fig. 3) due to sequence similarity.

Representative sequences of the studied samples showing variation in their nucleotides

TCS network showing three main haplotype groups (dash lines show the number of nucleotide substitutions)

Maximum likelihood phylogenetic tree of the studied individuals based on DNA sequences also produced three main clusters supporting the TCS network result (Fig. 4). Moreover, this phylogenetic tree showed that both case and control samples were mixed and were placed close to each other. Similarly, samples with different ethnic group backgrounds were also scattered throughout the phylogenetic tree. These results indicated that the sequences obtained from the queried SNP, as well as the neighboring sequences, were not associated with azoospermia. NCBI search for the sequence of present SNP (rs1462218557) showed that the following six known SNPs were in close vicinity of the queried SNP. They were rs555576178 INDEL (in-frame deletion), rs533994707 SNP (missense variant), rs577073776 SNP (missense variant), rs562944441 SNP (5′ UTR variant), rs574590943 SNP (5′ UTR variant), and rs541953116 SNP (5′ UTR variant). Therefore, sequence-based analyses contain rs1462218557 as well as these six SNPs.

Maximum likelihood tree of the studied individuals based on DNA sequences showing that the samples from both case (CA samples) and control (CO samples) are intermixed

We performed two types of association analyses to study association between different sequences of the studied samples and azoospermia. The LDA plot (Fig. 5) showed that there is no distinction between case and control samples and therefore DNA sequences do not differentiate the azoospermia patients from the controls. The Manhattan plot of LFMM analysis which is a Bayesian approach method also produced low p values for the studied DNA sequences showing no association between sequences and azoospermia (Fig. 6). Association between clinical features of the studied samples with geographical variables (Longitude and latitude) of their ethnic group also produced no significant association (p = 0.57, Fig. 7). This shows that though the clinical feature of the studied samples (particularly sperm characteristics which are absent in azoospermia individuals) differentiates these samples from each other, these features are not related to ethnic groups geographical features.

LDA plot showing no association between DNA sequences and azoospermia individuals. The control individuals (red colored) are not differentiated from azoospermia samples (blue colored samples)

Manhattan plot of LFMM analysis, showing non-significant p value for the studied sequences

RDA plot of clinical features and geographical variables of the studied individuals. (N = Longitude, E = Latitude)