Today, very large data sets of SNPs identified from whole genome sequencing are available, such as e.g. the huge resource provided by the UK Biobank (ukbiobank.ac.uk). Still, the data set analysed in our study possesses characteristics that make it advantageous for the SNP count analysis. In particular, all 1,222 individuals were ascertained and processed as a single cohort and therefore underwent identical methodology of variant calling, including the genotyping platform and sequence pre-processing, which allowed for minimising the technical bias of SNP calling. Moreover, this data set represents a timely and geographically uniform group of individuals of Polish origin (Kaja et al. 2022) and therefore excludes the ascertainment bias of SNP frequency due to population stratification and selection.
SNP distribution
In our study, we deliberately focused on SNP count instead of SNP density, even though the length of particular exons and introns varies considerably. However, while some studies have suggested a correlation between gene length or exon/intron count and SNP density (Lopes et al. 2021), the relationship is not always straightforward. Gene function, selective pressures, and genomic context can influence SNP counts (Deng et al. 2017). Therefore, introns and exons were in our study regarded as functional genomic units, and not as a mechanical sequence of nucleotides.
The functional role of the genomic region strongly determines the localization of polymorphisms, since SNPs in exons have a potential impact on gene products that on a further scope may be the cause of a disease or may alter quantitative phenotypes (Nair et al. 2021). Since introns exhibit various regulatory roles, the presence of polymorphism in introns may indirectly impact gene products or their expression levels (Mukherjee et al. 2018). Still, due to the generally more severe potential consequences of polymorphisms in exons, the expectation is that exons contain fewer SNPs than introns (Frigola et al. 2017), which was confirmed by our study. Moreover, genes with a low number of exons had the highest mean number of SNPs per exon, what was also observed in this study. This may be related to the fact that smaller genes are frequently expressed during an individual's lifetime because they are typically involved in functions that require fast responses, such as the immune system. These specific functions contribute to a higher variation which facilitates the response to and interaction with changing environment (Lopes et al. 2021).
SNP counts in introns
The biological role of introns is manyfold. They allow for alternative splicing (Bush et al. 2017) but additionally influence the stability of mRNA molecules (Gupta et al. 2013) and themselves contain noncoding RNA genes (Chorev and Carmel 2012) and regulatory elements, especially enhancers that affect the rate of transcription, known as the phenomenon of Intron Mediated Enhancement (Clancy and Hannah 2002; David-Assael et al. 2006). In our study, the significant excess of SNPs was observed in first introns. Of all the introns, the first one has been recognized as having special features and functions including, among others, correcting cytoplasmic localization of some mRNAs as well as transcriptional and translational regulation (Jo and Choi 2015). The important role of genetic variation in the first introns can also be anticipated by observing a very high number (over 120 since 2001) of publications reporting associations of SNPs located in the first intron with a variety of phenotypes measured in humans, animals, and plants (based on PubMed access on 10.01.2024). These specific roles may explain why the human DNA's first intronic sequence is considered the longest and highly dense of regulatory chromatin marks (Park et al. 2014; Jo and Choi 2019) Considering the criterion of SNP density, the above-mentioned studies identified the first introns as the most conserved regions, that however, provided a highly non-uniform distribution of SNPs along the intronic sequence (Majewski and Ott 2002) does not directly translate to the total low number of SNPs. Interestingly, in Arabidopsis thaliana, Back and Walther (2021) observed that the first introns harbour more SNPs than the subsequent introns.
SNP counts in exons
In the aforementioned study of Back and Walther (2021) that used Arabidopsis thaliana as the model genome, a high positive correlation was estimated between the sequence variation in the first exons and gene expression, which would explain higher genomic variability of the first exon observed in our study. Also, in a slightly different context of across-species comparison based on the reference genome sequence, Castle (2011) observed higher variability of coding regions in the proximity of the start and stop codons, so the ones typically corresponding to the first and the last exons, that is in line with our observation of SNP excess in the first and the last exons. Moreover, first and last exons include not only the protein-coding sequence but also the 5′- and 3′-untranslated regions (UTRs). The 5′UTR is the RNA sequence immediately upstream of the coding RNA. It is generally not a translated region but, due to the importance of the sequence for RNA transcription, stability, and translation, genetic variants modifying these elements are likely to have a profound effect. Analogously, the 3′UTR is located downstream of the coding sequence, and it is involved in regulatory processes, including RNA stability, mRNA translation, and localization. The 3′ UTR is characterized by binding sites for microRNAs and RNA-binding proteins and thus any variation of this region may lead to a change in gene expression (Steri et al. 2018). Although, UTRs are considered to have approximately the same genomic footprint as protein-coding regions still, polymorphisms within coding sequences may directly affect proteins and thus affect their function. This explains that higher SNP number in the first and last exons may be related to the presence of UTRs.