Computer Simulation of Biological Systems

This thesis investigates two biological systems using atomistic modelling and molecular dynamics simulation. The work is focused on: (a) the study of the interaction between a segment of a DNA molecule and a functionalized surface; (b) the dynamical modelling of protein tau, an intrinsically disordered protein. We briefly describe here the two problems; for their detailed introduction we refer respectively to chapter DNA and chapter TAU. The interest in the study of the adsorption of DNA on functionalized surfaces is related to the considerable effort that in recent years has been devoted in developing technologies for faster and cheaper genome sequencing. In order to sequence a DNA molecule, it has to be extracted from the cell where it is stored (e.g. the blood cells). As a consequence any genomic analysis requires a purification process in order to remove from the DNA molecule proteins, lipids and any other contaminants. The extraction and purification of DNA from biological samples is hence the first step towards an efficient and cheap genome sequencing. Using the chemical and physical properties of DNA it is possible to generate an attractive interaction between this macromolecule and a properly treated surface. Once positioned on the surface, the DNA can be more easily purified. In this work we set up a detailed molecular model of DNA interacting with a surface functionalized with amino silanes. The intent is to investigate the free energy of adsorption of small DNA oligomers as a function of the pH and ionic strength of the solution. The tau protein belongs to the category of Intrinsically Disordered Proteins (IDP), which in their native state do not have an average stable structure and fluctuate between many conformations. In its physiological state, tau protein helps nucleating and stabilizing the microtubules in the axons of the neurons. On the other hand, the same tau - in a pathological aggregation - is involved in the development of the Alzheimer disease. IDPs do not have a definite 3D structure, therefore their dynamical simulation cannot start from a known list of atomistic positions, like a protein data bank file. We first introduce a procedure to find an initial dynamical state for a generic IDP, and we apply it to the tau protein. We then analyze the dynamical properties of tau, like the propensity of residues to form temporary secondary structures like beta-sheets or alpha-helices