You are here

The Relationship Between DNA's Physical Properties and the DNA Molecule's Harmonic Signature, and Related Motion in Water--A Computational Investigation

Download pdf | Full Screen View

Date Issued:
2015
Abstract/Description:
This research investigates through computational methods whether the physical properties of DNA contribute to its harmonic signature, the uniqueness of that signature if present, and motion of the DNA molecule in water. When DNA is solvated in water at normal 'room temperature', it experiences a natural vibration due to the Brownian motion of the particles in the water colliding with the DNA. The null hypothesis is that there is no evidence to suggest a relationship between DNA's motion and strand length, while the alternative hypothesis is that there is evidence to suggest a relationship between DNA's vibrational motion and strand length. In a similar vein to the first hypothesis, a second hypothesis posits that DNA's vibrational motion may be dependent on strand content. The nature of this relationship, whether linear, exponential, logarithmic or non-continuous is not hypothesized by this research but will be discovered by testing if there is evidence to suggest a relationship between DNA's motion and strand length. The research also aims to discover whether the motion of DNA, when it varies by strand length and/or content, is sufficiently unique to allow that DNA to be identified in the absence of foreknowledge of the type of DNA that is present in a manner similar to a signature. If there is evidence to suggest that there is a uniqueness in DNA's vibrational motion under varying DNA strand content or length, then additional experimentation will be needed to determine whether these variances are unique across small changes as well as large changes, or large changes only. Finally, the question of whether it might be possible to identify a strand of unique DNA by base pair configuration solely from its vibrational signature, or if not, whether it might be possible to identify changes existing inside of a known DNA strand (such as a corruption, transposition or mutational error) is explored. Given the computational approach to this research, the NAMD simulation package (released by the Theoretical and Computational Biophysics Group at the University of Illinois at Urbana-Champaign) with the CHARMM force field would be the most appropriate set of tools for this investigation (Phillips et al., 2005), and will therefore be the toolset used in this research. For visualization and manipulation of model data, the VMD (Visual Molecular Dynamics) package will be employed. Further, these tools may be optimized and/or be aware of nucleic acid structures, and are free. These tools appear to be sufficient for this task, with validated fidelity of the simulation to provide vibrational and pressure profile data that could be analyzed; sufficient capabilities to do what is being asked of it; speed, so that runs can be done in a reasonable period of time (weeks versus months); and parallelizability, so that the tool could be run over a clustered network of computers dedicated to the task to increase the speed and capacity of the simulations. The computer cluster enabled analysis of 30,000 to 40,000 atom systems spending more than 410,000 CPU computational hours of hundreds of nano second duration, experimental runs each sampled 500,000 times with two-femtosecond (")frames.(")Using Fourier transforms of run pressure readings into frequencies, the simulation investigation could not reject the null hypotheses that the frequencies observed in the system runs are independent on the DNA strand length or content being studied. To be clear, frequency variations were present in the in silicon replications of the DNA in ionized solutions, but we were unable to conclude that those variations were not due to other system factors. There were several tests employed to determine alternative factors that caused these variations. Chief among the factors is the possibility that the water box itself is the source of a large amount of vibrational noise that makes it difficult or impossible with the tools that we had at our disposal to isolate any signals emitted by the DNA strands. Assuming the water-box itself was a source of large amounts of vibrational noise, an emergent hypothesis was generated and additional post-hoc testing was undertaken to attempt to isolate and then filter the water box noise from the rest of the system frequencies. With conclusive results we found that the water box is responsible for the majority of the signals being recorded, resulting in very low signal amplitudes from the DNA molecules themselves. Using these low signal amplitudes being emitted by the DNA, we could not be conclusively uniquely associate either DNA length or content with the remaining observed frequencies. A brief look at a future possible isolation technique, wavelet analysis, was conducted. Finally, because these results are dependent on the tools at our disposal and hence by no means conclusive, suggestions for future research to expand on and further test these hypothesis are made in the final chapter.
Title: The Relationship Between DNA's Physical Properties and the DNA Molecule's Harmonic Signature, and Related Motion in Water--A Computational Investigation.
60 views
24 downloads
Name(s): Boyer, Victor, Author
Proctor, Michael, Committee Chair
Thompson, William, Committee Member
Karwowski, Waldemar, Committee Member
Calloway, Richard, Committee Member
University of Central Florida, Degree Grantor
Type of Resource: text
Date Issued: 2015
Publisher: University of Central Florida
Language(s): English
Abstract/Description: This research investigates through computational methods whether the physical properties of DNA contribute to its harmonic signature, the uniqueness of that signature if present, and motion of the DNA molecule in water. When DNA is solvated in water at normal 'room temperature', it experiences a natural vibration due to the Brownian motion of the particles in the water colliding with the DNA. The null hypothesis is that there is no evidence to suggest a relationship between DNA's motion and strand length, while the alternative hypothesis is that there is evidence to suggest a relationship between DNA's vibrational motion and strand length. In a similar vein to the first hypothesis, a second hypothesis posits that DNA's vibrational motion may be dependent on strand content. The nature of this relationship, whether linear, exponential, logarithmic or non-continuous is not hypothesized by this research but will be discovered by testing if there is evidence to suggest a relationship between DNA's motion and strand length. The research also aims to discover whether the motion of DNA, when it varies by strand length and/or content, is sufficiently unique to allow that DNA to be identified in the absence of foreknowledge of the type of DNA that is present in a manner similar to a signature. If there is evidence to suggest that there is a uniqueness in DNA's vibrational motion under varying DNA strand content or length, then additional experimentation will be needed to determine whether these variances are unique across small changes as well as large changes, or large changes only. Finally, the question of whether it might be possible to identify a strand of unique DNA by base pair configuration solely from its vibrational signature, or if not, whether it might be possible to identify changes existing inside of a known DNA strand (such as a corruption, transposition or mutational error) is explored. Given the computational approach to this research, the NAMD simulation package (released by the Theoretical and Computational Biophysics Group at the University of Illinois at Urbana-Champaign) with the CHARMM force field would be the most appropriate set of tools for this investigation (Phillips et al., 2005), and will therefore be the toolset used in this research. For visualization and manipulation of model data, the VMD (Visual Molecular Dynamics) package will be employed. Further, these tools may be optimized and/or be aware of nucleic acid structures, and are free. These tools appear to be sufficient for this task, with validated fidelity of the simulation to provide vibrational and pressure profile data that could be analyzed; sufficient capabilities to do what is being asked of it; speed, so that runs can be done in a reasonable period of time (weeks versus months); and parallelizability, so that the tool could be run over a clustered network of computers dedicated to the task to increase the speed and capacity of the simulations. The computer cluster enabled analysis of 30,000 to 40,000 atom systems spending more than 410,000 CPU computational hours of hundreds of nano second duration, experimental runs each sampled 500,000 times with two-femtosecond (")frames.(")Using Fourier transforms of run pressure readings into frequencies, the simulation investigation could not reject the null hypotheses that the frequencies observed in the system runs are independent on the DNA strand length or content being studied. To be clear, frequency variations were present in the in silicon replications of the DNA in ionized solutions, but we were unable to conclude that those variations were not due to other system factors. There were several tests employed to determine alternative factors that caused these variations. Chief among the factors is the possibility that the water box itself is the source of a large amount of vibrational noise that makes it difficult or impossible with the tools that we had at our disposal to isolate any signals emitted by the DNA strands. Assuming the water-box itself was a source of large amounts of vibrational noise, an emergent hypothesis was generated and additional post-hoc testing was undertaken to attempt to isolate and then filter the water box noise from the rest of the system frequencies. With conclusive results we found that the water box is responsible for the majority of the signals being recorded, resulting in very low signal amplitudes from the DNA molecules themselves. Using these low signal amplitudes being emitted by the DNA, we could not be conclusively uniquely associate either DNA length or content with the remaining observed frequencies. A brief look at a future possible isolation technique, wavelet analysis, was conducted. Finally, because these results are dependent on the tools at our disposal and hence by no means conclusive, suggestions for future research to expand on and further test these hypothesis are made in the final chapter.
Identifier: CFE0005930 (IID), ucf:50835 (fedora)
Note(s): 2015-12-01
Ph.D.
Engineering and Computer Science, Industrial Engineering and Management Systems
Doctoral
This record was generated from author submitted information.
Subject(s): NA -- resonance -- molecular vibration -- molecular motion -- molecular simulation
Persistent Link to This Record: http://purl.flvc.org/ucf/fd/CFE0005930
Restrictions on Access: campus 2020-12-15
Host Institution: UCF

In Collections