The Breakdown of Symmetry in Word Pairs in 1,092 Human Genomes

Single strand symmetry has been observed in several genomes, and some authors have associated this phenomenon to genome evolution. However, it is still not clear how strong and exceptional this phenomenon is. We use next-generation sequencing data from a sample of 1,092 human individuals made available by the 1000 Genomes Project. To evaluate the phenomenon of symmetry of single-strand human genomic DNA, we explore and analyze these 1,092 human genomes and 1,092 randomly generated sequences, each forced to mimic the nucleotide frequency distribution of their real counterpart. Our methodology is based on measurements, traditional and equivalence statistical tests using different parameters. By statistical testing we find that the global symmetries phenomenon is significant for word lengths  smaller than 8. When we evaluate the global symmetry scores, we obtain strong values for all word lengths and both types of sequences under study. However, the symmetry scores in human genomes reach higher values and have lower dispersion than those in random sequences. We also find that human and random symmetry scores are significantly different. We conclude that in the human genome, the differences between symmetric words are higher than in random sequences, but the correlation between symmetric words in human genomes is higher.