Algorithms for (i) Prediction of Secondary Structure of Nucleic Acids from CD Spectra using Machine Learning and (ii) Determination of Dissociation Constant Associated with Biomolecular Interactions

Vinothini, and Rathinavelan, Thenmalarchelvi (2019) Algorithms for (i) Prediction of Secondary Structure of Nucleic Acids from CD Spectra using Machine Learning and (ii) Determination of Dissociation Constant Associated with Biomolecular Interactions. Masters thesis, Indian institute of technology Hyderabad.

Text
Mtech_Thesis_TD1419_2019.pdf
Restricted to Repository staff only until 28 June 2024.
Download (2MB) | Request a copy

Abstract

Circular Dichroism (CD) has been profoundly used to gain secondary structural information about both protein/nucleic acid structures. Several algorithms have been evolved to quantitatively estimate the secondary structural content of proteins from CD data, but it is not the case for nucleic acids. There is relatively limited success in correlating the observed CD spectrum with structural parameters of nucleic acids. To achieve this, a reference library enclosing the CD spectra of known forms of nucleic acids must be created. The information gained out of the library mentioned above can be exploited to anticipate the unknown structure by fitting its measured CD spectrum with the reference set. The present investigation checks the feasibility of using machine learning algorithms to predict the secondary structure of nucleic acids from CD spectra. Several experimental methods are available to determine the dissociation constant for macromolecule-ligand interaction. One such method is CD spectroscopy, but, requires manual pre-processing of data points. Curve fitting of data points yields binding curve and parameters like, dissociation constant. An algorithm to perform curve fitting in addition to pre-processing of CD data would simplify the entire process. This study focusses on developing an algorithm that employs R programming to estimate the dissociation constant. This is method considers cooperativity and fits the data points in Hill equation, and for non-cooperative binding studies, the data points are fitted in the Law of Mass Action equation.

[error in script]

IITH Creators: