Search

We have developed a simple algorithm of counting the probability of each possible type of di-nucleotide occurring in a DNA sequence to determining homologs and phylogeny between mammals. A 2-dimensional representation was developed for visualization. This method does not require any DNA alignment or for the DNA sequences to be of the same length. We applied our tools to determining homologs between related zinc finger structures, as well as determining mammalian phylogeny. We also demonstrate this method's ability to classifying dissimilar DNA sequences. Likely future work along these lines will be to increase our vector to include dimensions for other variables such as fractal dimension, large automated DNA searches, and segmenting the DNA sequence before calculating our vector in order to keep some position data.

A C2H2 zinc finger nucleotide sequence can be expressed as a numerical sequence when each nucleotide is assigned its proton number. A resulting zinc finger numerical sequence could be investigated for its fractal dimension in terms of evolution, multiple finger structure flexibility, and 3-finger structure specificity. The ZNF714 sequence was found to have a fractal dimension higher than its evolutionary ancestor ZNF431, consistent with the recently reported evolutionary direction using traditional analysis. The ZNF91 sequence has 36 zinc fingers and their fractal dimensions form a Gaussian-like histogram, suggesting a maximal flexibility for capacity dimension. The CpG di-nucleotide percentage in the three zinc fingers of EGR1 is found to correlate with the fractal dimension while the five zinc fingers in ZFP91 separated into human-chimp and mouse-rat-dog groups. Extension of the fractal investigation to short genetic sequences is discussed.

We have developed a fractal algorithm to supplement the current analysis in lung-CT scans on non-solid pulmonary nodule. The Cornell University Lung-CT Scans database and the US National Institute of Health ImageJ software were used. We found that the inner core average intensity could be a marker for the fractal dimension of a non-solid nodule. Z-scan image analysis of the adjacent boundaries showed that the fractal dimension has a range of about 0.1 in the studied cases. Extension of this proof-of-concept fractal analysis to growth rate estimation is also discussed.

The CpG di-nucleotide percentage and exon fractal dimension fluctuation were investigated with respect to the recently identified Marfan syndrome exons in the FBN1 gene. The CpG di-nucleotide percentage was found to rank high in exon 1 (9%), exon 44 (6.5%), exon 24 (5.3%) and exon 27 (4.8%). The most significant Marfan exon group was reported to be exon 24- 32. An exon nucleotide sequence can be expressed as a numerical sequence when each nucleotide is assigned its proton number. The resulting exon nucleotide numerical sequence could be investigated for its fractal dimension. The fractal dimension exhibited fluctuation across the 65 FBN1 exons and the values formed a Gaussian-like distribution (N = 65). The moving average fractal dimension was found to exhibit a broad peak around exons 24 to 32, consistent with the reported Marfan syndrome exon group.

The fluorescent speckle microscope pattern of newt lung epithelial cell filamentous actin during migration was analyzed using the Higuchi fractal method. The public domain datasets of 2004 PNAS Vallotton et al. were studied. The time series data from time lapsed images exhibited similar fractal dimensions (about 1.68 to 1.82) for various cell regions. Computer simulation of random time series data suggested an average fractal dimension of about 2 with a standard deviation of about 0.027. Speckle trend removals revealed pulsation features which were further studied using a sinusoidal signal model commonly used in gene regulatory studies. Gaussian noise models mixed with sinusoidal signal were used to simulate the observed fractal dimensions. The extracted constraints could be interpreted as polymerization and control pathway related. The administration of contractility promotion drug calyculin-A resulted in an increase of the overall image fractal dimension, which is consistent with the loss of correlation due to random direction contractions. It appears that high fractal dimension could be interpreted as being due to further randomization. Two doses of 20 nM calyculin-A increased cell noise by about 45%. Extension of this proof-of-concept fractal analysis to clinical treatment was also discussed.