Master thesis proposals

Here are summarised master thesis projects offered in the VARIAMOLS and SBP group.

For further information and discussion please write to raffaello.potestio [at] unitn.it

Project A | Implementation and validation of a user-friendly interface for Hamiltonian adaptive resolution (H-AdResS) simulations in LAMMPS

Background. One of the most challenging issues in molecular simulations is to match the gap in the broad range of length and time scales which characterises soft matter systems. One method which proved capable of reducing this gap is the Hamiltonian adaptive resolution simulation (H-AdResS) scheme, which enables the concurrent usage, in the same simulation setup, of two models representing the same system at different resolution. By these means, an accurate but computationally costly model (e.g. all-atom) is employed in a small subdomain (e.g. for a small protein or part of it), while a simplified and more efficient low-resolution representation is used for the rest (w.g. the solvent molecules). This method is currently implemented in the LAMMPS simulation software.

Objective. The master thesis work will focus on the development of an efficient, user-friendly interface to setup the system and the simulation. This also includes the integration with general-purpose software for the manipulation of molecular simulation data (e.g. VMD). The whole computational setup will be validate on scientifically relevant problems, such as the simulation of macromolecules in adaptive resolution solvent in equilibrium and out of equilibrium.

Perks. The project is framed in a collaboration with Prof. Davide Donadio at the University of California Davis, USA, and Dr. Robinson Cortes-Huerto at the Max Planck Institute for Polymer Research in Mainz, Germany. The possibility of short (1 to 3 months) missions abroad is thus foreseen.

Project B | Dual-resolution simulations of membrane proteins embedded in lipid bilayers

Background. Lipid membranes are a crucial element of the cell, constituting its walls that contain the cellular machinery, allowing compartimentalisation, shielding the DNA (at least in eukaryotic cells) and regulating the communication between the inside and the outside. The importance of the role played by lipid membranes in the cell as well as in several other biological contexts can hardly be overestimated, and this prominence has pushed the development of various models and methods to investigate their properties and behaviour in silico. Particularly relevant is the availability of all-atom models for detailed simulations, and simplified, coarse-grained representations to reach the size and length scales of large assemblies of hundreds of thousands of lipids. Unfortunately, a large number of scientifically interesting and important questions cannot be tackled due to the fact that a high degree of detail is often required in simulations of systems too large to be performed with standard all-atom models.

Objective. The master thesis work aims at performing dual-resolution simulations of a lipid membrane mostly described by the coarse-grained MARTINI force field in combination with a high-resolution treatment of a small region including an atomistic model of membrane, solvent and a membrane protein. The system will be implemented in the LAMMPS simulation software, and will enable the detailed investigation of the function of membrane proteins in large lipid bilayers, especially tackling the role of membrane fluctuations.

Perks. The project is framed in a collaboration with Prof. Alberto Giacomello at the University of Rome La Sapienza.

Project C | Development and validation of a hybrid all-atom/coarse-grained model of DNA

Background. The comprehension of the properties of genetic material such as DNA and RNA is a major scientific challenge. Computational methods such as molecular dynamics simulations are proving of ever increasing importance to achieve an atomic-scale understanding of how these macromolecules work and carry out their pivotal function for the sustenance of life. However, the size of DNA molecules, which can reach the hundreds of millions of base pairs in chromosomes, makes the problem unachievable by means of high-resolution, all-atom models. Low-resolution representations, such as coarse-grained models, enable researcher to reach biologically relevant time and length scales with limited computational effort. These models, though, lack the fine-grained detail which is often necessary to acquire information about the local mechanical and dynamical properties of DNA.

Objective. The aim of this master thesis work is to develop and validate a dual-resolution model of DNA which combines a standard, all-atom representation and the coarse-grained OxDNA force field. The resulting model will be tested on scientifically relevant case studies and, wherever possible, validated against experimental results.

Perks. The student will be part of a multidisciplinary team involving a collaboration with the Ca' Foscari University of Venice and will work closely with students and postdocs of both Universities.

Project D | Characterisation of the entropic landscape of a polymer with adhesive beads

Background. Polymers are long molecules constituted by small units connected one after the other in a linear chain. Many relevant molecules are polymers, such as DNA, RNA, proteins, as well as several artificial ones, plastic and rubber being among the most prominent examples. Single polymer chains can very likely be found in a self-entangled conformation, that is, the configurations that a polymer can attain at equilibrium will most likely contain one or more knots, just as in shoelaces or other cords. The knot spectrum of simple polymers has been thoroughly studied and is currently well understood, however the properties of knotted polymers become much more complex as soon as a few nontrivial interactions are allowed among the molecule’s constituents.

Objective. The goal of this master thesis work is to characterise the knotting properties of polymer chains featuring various patterns of simple attractive interactions. The investigation, relying on molecular dynamics simulations, will aim at describing the knotting process in terms of the conformational entropy of the system. This work bears great importance for the understanding of the knotting process in biological heteropolymers (e.g. proteins) and for the design of artificial molecules with tailored topological properties.

Perks. This work will be carried out in collaboration with Dr. Luca Tubiana at the Phys. Dept. of UniTn, and will offer the opportunity to interact with prestigious European institutions within the COST Action EUTOPIA - European Topology Interdisciplinary Action.

Project E | Sequence-property relation of lattice polymers studied through the zeroes of the partition function

Background. In spite of their simplicity, lattice models of polymers are capable of providing a substantial qualitative and even quantitative insight in the physics of real polymeric macromolecules. A still open problem is the design of polymers featuring desired characteristics, that is, to define a sequence of monomers with known chemical properties such that the whole macromolecule displays a specific (structural or thermodynamical) behaviour.

Objective. This project aims at addressing this problem in general terms, establishing a link between the sequence of the polymer and the resulting thermodynamics. This analysis will be carried out making use of the well-known relationship existing between the location of zeroes of a system’s partition function and its thermodynamic properties.

Perks. This project is framed in the context of the VARIAMOLS ERC project; the student will have access to the computational resources of the research group.

Project F | Mapping entropy and maximally informative samples

Background. One of the most relevant and less studied aspects of coarse-graining is the issue of the mapping, that is, the quantitative relationship which defines the large-scale, collective degrees of freedom of a lower-resolution model in terms of the coordinates of the underlying, fine-grained high-resolution particles. The mapping impacts in particular the entropic component of the effective potential, or multi-body of mean force, since it accounts for the number of the original system’s microstates which map onto one of the coarse-grained system’s macrostates. The identification of an optimal mapping thus requires that the amount of entropy lost in the coarse-graining procedure is minimal.

Objective. This project aims at investigating the aforementioned issue in terms of maximally informative samples theory, a powerful framework which enables one to quantify the statistical properties of a sample of instances grouped according to a given protocol. The systems will be investigated also in the light of more standard tools of coarse-graining, namely the relative entropy approach, which is commonly employed to parametrise effective models by minimising a measure of information loss.

Perks. This project is framed in the context of the VARIAMOLS ERC project; the student will have access to the computational resources of the research group.