The effects of C-G content and mutations on the Fourier transform method for periodicity

It is well-known that the Fourier spectrum of a DNA protein-coding region exhibits an f = 1/3 peak. This is due to an unbalanced nucleotide distribution and open reading frame (ORF) positional bias that introduces a 3-base periodicity into the sequence. Until now, the f = 1/3 property has mainly been used to detect protein-coding regions, but in our paper, we use the f = 1/3 spectral height to detect the strength of the periodicity in these regions. First, we quantify the effects of sequence length, nucleotide distribution, ORF positional bias, and mutations on this spectral measure. Then once we know how the spectral peak height changes in response to these aberrations, we investigate a sliding window technique in an attempt to detect changes in CG content. Detection of frameshift mutations are important for finding disease, and it has been hypothesized that nucleotide bias variation signifies a change in the function of encoded proteins.