Ancestral genomes and population sequencing data reveal strand- and context-specificity of polymorphic G4 sites

Recent studies highlight the important functional roles of non-canonical DNA conformations. One of such structures, G-quadruplex (G4), was shown to be involved in multiple processes within a cell such as telomere maintenance, gene regulation, protein translation and alternative splicing. On the other hand, DNA in non-double stranded context can hinder replication and repair processes and, indeed, a higher rate of polymorphisms was reported at G4 sites. However, strand-specificity, positional and nucleotide contexts of nucleotide substitutions at G4 sites are under-investigated. Here, we combine ancestral genome data and DNA variants from large populational sequencing project to investigate substitution patterns within putative DNA quadruplexes. We confirm the overall elevated rate of base replacement except transitions at CpG sites, which are less likely than in the rest of the genome potentially due to hypomethylated status of G4s. Within G4 loops, there is a trend of replacing weak bases (A, T) with strong bases (G, C) that could promote DNA duplex stabilization. However, this trend is less pronounced when derived allele is rare in the human population. The most over-represented thymine to guanine replacement is about order of magnitude more frequent within G4 than in non-G4 regions. The analysis of nucleotide context of the substitutions shows clear difference between G-rich and C-rich of DNA quadruplexes implying that the strands of the quadruplex might have different mutation dynamics. Finally, we show that the observed deviations from random mutation accumulation result in a biased nucleotide composition of G4 loops rich in adenines. Future studies should reveal specific mutation and selection processes that shape the content of G4-associated DNA polymorphisms.