Nucleotide Composition of Ultra-Conserved Elements Shows Excess of GpC and Depletion of GG and CC Dinucleotides

Abstract

The public UCNEbase database, comprising 4273 human ultra-conserved noncoding elements (UCNEs), was thoroughly investigated with the aim to find any nucleotide signals or motifs that have made these DNA sequences practically unchanged over three hundred million years of evolution. Each UCNE comprises over 200 nucleotides and has at least 95% identity between humans and chickens. A total of 31,046 SNPs were found within the UCNE database. We demonstrated that every human has over 300 mutations within 4273 UCNEs. No association of UCNEs with non-coding RNAs, nor preference of a particular meiotic recombination rate within them were found. No sequence motifs associated with UCNEs nor their flanking regions have been found. However, we demonstrated that UCNEs have strong nucleotide and dinucleotide sequence abnormalities compared to genome averages. Specifically, UCNEs are depleted for CC and GG dinucleotides, while GC dinucleotides are in excess of 28%. Importantly, GC dinucleotides have extraordinarily strong stacking free-energy inside the DNA helix and unique resistance to dissociation. Based on the adjacent nucleotide stacking abnormalities within UCNEs, we conjecture that peculiarities in dinucleotide distribution within UCNEs may create unique 3D conformation and specificity to bind proteins. We also discuss the strange dynamics of multiple SNPs inside UCNEs and reasons why these sequences are extraordinarily conserved.

Keywords: DNA structure; bioinformatics; computational biology; evolution; genetic variation; genomics; polymorphism.

Figure 1

Distribution of SNP relative frequencies by their alternative allele abundance inside UCNEs and the whole genome. This is a graphical representation of data from Table 1 for the second up to fiftieth bins for columns 3 and 5. Starting from the second bin, the relative frequency of SNPs inside the whole genome is always higher than inside UCNE sequences, and the difference becomes more dramatic with the increase of alternative allele frequency (bin consecutive order).

Figure 2

Number of alternative alleles with the frequencies up to 50% inside the 4271 UCNE sequences among 2504 individuals from five regions. Individuals are represented in five groups, depending on their ethnicity and according to their classification in the 1000 Genomes Database. AFR represents African populations (navy blue), AMR—Americans populations (red), EAS—East Asian (yellow), EUR—Europeans (blue), and SAS—South Asia (green). Each individual is represented by a colored bar, and its position along horizontal axis corresponds to the total number of alternative alleles inside the UCNEs in each person.

Figure 3

Distribution of meiotic recombination rates inside UCNEs versus random genomic positions (so-called “random-UCNEs”). Recombination rates were divided into equal-sized intervals of 0.02 centimorgans (cM) per one million nucleotides, which are shown along the horizontal axis. The number of UCNE and “random-UCNE” sequences that have a recombination rate within a particular interval (bin) are plotted along the vertical axis.