Abstract
Calculating peak-height ratios between single-nucleotide polymorphisms (SNP) alleles in sequencing chromatograms is a practical method for estimating their copy number proportions (CNPs). However, it is surprising that sequencing DNA from different directions might yield different results. We analyzed three adjacent SNPs within the ovine period circadian-clock 2 (PER2) gene that displayed such behavior. We compared Sanger and DNA-seq sequencing for this locus and applied high-resolution melt and MFOLD analyses to point to the DNA secondary structure that underlined this phenomenon. A synthetic system of oligonucleotides cloned into plasmids was used to further test the effect of such structures on sequencing. Our analyses indicated that a stem-loop structure capable of G-T pairing mediated the orientation bias by stabilizing this structure for specific alleles in heterozygous situations. We propose that this wobble-like pairing hinders DNA polymerase passage on one strand while, on the complementary strand, the nonpaired A-C nucleotide counterparts allow unobstructed replication. Experimentation with synthetic amplicons that form similar stem-loop structures supported our hypothesis. We coined the term 'replication diode' for this effect and demonstrated that we can minimize it by lowering DNA and salt concentration. We also demonstrated that common genomic palindromes might induce the replication diode effect by applying bidirectional sequencing to an amplicon containing the palindrome within the human miRNA 1-1 gene. Hence, to obtain reliable peak-height ratios, bidirectional sequencing should be practiced at the lowest possible ionic strength whenever estimating CNPs. Further research is needed to determine whether the observed variable stem-loop structures affect PER2 regulation in vivo.