Performing computational molecular dynamics (MD) simulations of small-molecule systems has become one of the most prominently used methods in studies of molecular structure and behavior. However, MD is a computationally expensive and time-consuming methodology because of the requirement of computing detailed interactions among atom-atom pairs. There is great interest, therefore, in reducing the time and computational power needed to approximate real-world systems. Most commonly, such efforts have employed machine learning techniques to predict extensive properties of molecular systems. Here, we propose accelerating simulations by predicting conformational changes - a prospect that has not yet been fully explored. Previous work attempted applying a linear dynamical analysis algorithm named Dynamic Mode Decomposition to MD data, which has been shown to be ineffective through a multiresolution analysis. We propose herein the use of Sparse Identification of Nonlinear Dynamical Systems (SINDy), a nonlinear model which has been shown to accurately decipher the governing equations of dynamical systems. We will be testing the effectiveness of SINDy with MD data by performing an iterative error analysis while varying the initial parameters of the dataset, thereby gaining a better understanding of how much data (and in what form) should be inputted to maximize the accuracy of a simulated SINDy model of an MD dataset. If shown to be sufficiently accurate, we then can implement SINDy simultaneously with MD in an active learning loop to save time and computational power while maintaining a high degree of predictive capability for peptide conformations. The current goal is to obtain a deeper understanding of peptide conformational changes that could, in the future, be combined with machine learning techniques to greatly accelerate classical MD simulations.

This project is supported by the UW Computational Neuroscience Center, and the DMREF Program of NSF through the MGI platform under DMR# 1629071, 1848911, and 1922020.

Reactive astrogliosis is a condition where astrocytes, a type of brain cell, undergo morphological – shape - changes upon exposure to brain injury. Morphological changes of astrocytes are a key indicator of activation and can be beneficial in stopping initial brain injury effects, but chronic activation can drive glial scarring, which is detrimental for full recovery of normal brain function. Glial scarring has been linked to several diseases, including traumatic brain injury, Alzheimer’s Disease, and dementia. The purpose of this work is to quantitatively analyze the relationship between frequency and extent of reactive astrogliosis with relation to distance from the primary site of brain injury. My approach is to build a Python-based image analysis pipeline to quantify astrocyte cell features. The Nance Lab’s prior work using Python packages was effective in developing a pipeline to identify and quantify microglial - a different type of brain cell - shape properties. We are now building a pipeline to study astrocytes in brain slices from the injured preterm ferret brain, which were stained with an antibody for glial fibrillary acidic protein (GFAP). Images of cells at 20x magnification are provided by Dr. Tommy Wood. Since response to injury can be brain region and animal sex dependent, I analyze astrocyte cell features in each region of the brain from both sexes of ferrets. I used SciKit-Image along with other packages to segment, label, and quantify features of our cells, including perimeter, area, and circularity, among others. Expected results include a significant reduction in area and an increase in perimeter (larger surface-to-volume ratio) of cells that are closer to the injury. This image analysis pipeline will give us quantitative information about the cells morphology which are associated with biological markers that can be targets for future therapeutic treatment. Clear biological markers help researchers develop better treatment.

The goal of this project is to develop a dynamically evolving connectionist model that more closely resembles the brain through its information-processing. Over the years, AI has shifted from the first generation of feedforward systems to the use of recurrent or convolutional Neural Networks. The third and newest generation of AI models, the brain-based models, and the Spiking Neural Network (SNN), attempts to bridge the gap between Neuroscience and ML using biologically realistic models like Θ-model, LIF, Izhikevich, HR, HH. These models, however, are still a black box leaving very little control or understanding on the learning process within the system without the access to the inner structure of the network. In addition, these systems are highly inefficient, slow, and very complex due to the limitations imposed by the hardware and explicit simulation of partial differential equations. Real world problems require “flexible learning and dynamically adaptive connectionist systems” that are capable to adapt and accommodate new input in real time. Current solutions have focused on varying the weights within a system rather than focusing on how connections within the system are formed. Based on our understanding from organismal brain structures, our approach, called biomimetic information codec, .bic, is a morphologically-adaptive coding hierarchical network that form in accordance with energy minimization - driven by dissipation of "heat" generated by the training data - constructing cortices and connectome for processing of information. My first objective herein is to quantitatively compare detailed structures between biological (fly brain) and .bic. networks using a random matrix approach.

Engineered solid binding peptides can be used as molecular tools for a variety of bio/nanotechnology applications, especially in interfacing biology with solid-state devices at bio/nano soft interfaces. The control of surface organization, and therefore peptide-solid interactions, is critical and involves surface phenomena such as binding, surface diffusion, and self-organization on atomically flat solids. Each of these phenomena requires the knowledge of peptide’s folding patterns which are, however, difficult to study both experimentally and computationally. Molecular dynamics, MD, has been used to computationally model peptide/solid interactions, but without information regarding the energy landscape of peptide conformations the challenge of predictive design remains. While several methods exist for finding the energy landscapes of single peptide systems, currently no approach handles multi-peptide/surface systems. Here we use Time-Varying Autoregression with Low Rank Tensors, TVART, to efficiently explore the energy landscapes of such systems, aiming to find accurate linear approximations for predictive design of peptides at bio/nano interfaces. Using TVART, with each slice representing a discrete time window, allows for temporal smoothness and high predictive accuracy. It is anticipated that some descriptions of conformation will be better suited to describe peptide conformation energy landscapes than others; based on this premise, we examined interatomic distances/adjacencies and peptide backbone torsion angles as descriptions of peptide conformation. Through such analyses, it is becoming possible to describe how peptide conformations in multi-peptide/surface systems evolve through the energy landscape and settle into energy minima (stable conformations). These conformations can then be corroborated with experimental validation of peptide self-organization on the surface using scanning probe microscopy techniques with sub-A resolutions. The combination of computational modeling and high-resolution experiments is expected to aid predictive design platforms for future applications in biosensors, bioelectronics, and logic devices.

Brain extracellular matrix (ECM) structure mediates many aspects of neuronal function. When ECM structure becomes dysregulated in neurological disease, one resulting impact is impaired neuronal function. Therefore, probing changes in the ECM structure could provide insights into disease mechanisms and expose potential therapeutic pathways. Previous work in our group determined that degrading neural ECM structures leads to a significant increase in the diffusive ability of nanoparticles navigating the brain extracellular space. However, this diffusion-based analysis provides little insight into changes in ECM-specific morphology or structure; analysis only predicts if they are present and the degree of alteration from normal. With this project, we aimed to characterize changes in perineuronal net (PNN) structure with high spatial resolution using a fluorescence-based imaging approach. PNNs, a structure that impact neuronal function were stained using a fluorescently labeled lectin (Wisteria floribunda agglutinin) and images acquired via confocal microscopy. Images were collected from the cortex of brains spanning an age range of post-natal day 14 to adulthood. We first manually quantified morphological features associated with PNNs, including the total number of branches, average branch length, and total branch length using the image processing program ImageJ. To reduce image processing time and minimize user-bias, I am building a Python-based, automated quantification workflow for future use. Regarding the manually quantified PNN features, we observed an increase in both the total branch length/PNN and average branch length/PNN as brains increased in age. The total number of branches/PNN remained relatively consistent across age groups. Future work will focus on applying the same approach to a model of neonatal hypoxia ischemia to study injury effect on PNN structure. Collectively, this project established a methodology that can be applied for enhanced characterization of ECM-related structural changes induced by neurological disease, and has the potential to unveil new avenues of therapeutic intervention.

Biological organisms are capable of generating inorganic materials to fulfil their needs in a process called biomineralization, and have been shown to exhibit control over the properties, such as size or shape, of the materials. Biomineralization occurs at mild conditions, which has generated a growing interest in replicating this process in a lab environment and developing methods that would allow researchers to apply this technique of biomimicry to the synthesis of inorganic materials that are not directly created through biomineralization. Biomimetic synthesis involves the use of biomolecules, such as proteins, DNA, or peptides, as templates or to induce mineralization. The mechanisms involved in synthesizing titanium dioxide, TiO2, using this technique of biomimicry are the specific interest of this project. Titanium dioxide is an oxide commonly found in titanium-containing ores, and has wide-spread application such as in pigments, pharmaceuticals, cosmetics, and photocatalysis. My work within this project specifically aims to understand the thermodynamics of the equilibrium equation between titanium(IV) bis(ammonium lactate)dihydroxide (TiBALDH), the preferred reagent for biomimetic titanium dioxide, and titanium dioxide. I created optimized structures of each reactant and product, and then I ran density functional theory (DFT) calculations using the computational chemistry software Gaussian to understand the energetics of each molecule. My use of these calculations to understand the energetics and mechanisms of this titanium dioxide reaction, as well as similar reactions, will help inform methods of improving the synthesis of titanium dioxide and other inorganic materials.

The self-assembly behavior of peptides on inorganic surfaces such as highly oriented pyrolytic graphite (HOPG) remains a highly studied field with numerous applications. Some of these applications include the integration of engineered devices for biomedical purposes. Overall, it is important to understand the types of interactions occurring at protein-surface interfaces, and which of these interactions govern the primary self-assembly behavior of the molecules. A number of experimental methods may be employed to better understand the behavior at these interfaces. Atomic force microscopy (AFM) is a well-established method, and was used to image self-assembly samples of charged peptides on graphite. Fixed concentrations of negatively charged peptides were placed in varying concentrations of KCl solutions to determine the effect of salt concentration on surface coverage and ordering of peptides on HOPG. Additionally, other factors were investigated, including the effect of step edges and incubation time on ordering. AFM images of the peptides were analyzed using Gwyddion to extract the surface coverage and degree of ordering. With increasing salt concentration, it was found that peptide assembly on graphite generally increased in KCl ranges of 1μM to 1mM. Additionally, the degree of peptide ordering increased with this range of salt concentration, displaying that charged peptides are able to pack closer together in the presence of an electrolyte. The electrostatic screening effect can explain this close-packing behavior, in which positively charged K+ ions are able to surround a negatively charged peptide and decrease the effective coulombic force between any two adjacent peptides. Overall, several factors may influence the assembly and ordering of peptides on HOPG, including electrolyte concentration, density of step-edges in the graphite, and charge of the peptide in question. This research has broader implications in the bio-medicial field with applications to engineered devices, and may prove important for future research.

The study of cell morphology is important and prevalent in understanding normal and pathological conditions in the brain. Brain cells are common targets for treatment for brain diseases. Specifically, microglia – the brain’s resident immune cells – undergo a range of morphological changes in response to injury and are targets of many mitigating treatments. Image processing has been a valuable tool to assess microglial cell morphology via the analysis of microglial shape features and there continue to be additional opportunities for further investigation. Prior research has indicated a connection between features such as solidity and extent, two shape features that measure the ratio of cell areas. In this study, we examine three shape features of fluorescently labeled microglia: Euler number, extent, and solidity, in the context of ischemic injury. Ischemic injury was modeled using oxygen-glucose deprivation (OGD) in cultured whole hemisphere brain slices. Using python, images were thresholded with the Otsu threshold. Shape features were extracted from the binarized images. These shape features were analyzed based on brain region (cortex, hippocampus, thalamus), generalized treatment type (non-treated, injured, injured with treatment), and specific treatment type (OGD 0.5 hour, 1.5 hours, 3 hours, 1.5 hours with azithromycin treatment, 3 hours with superoxide dismutase treatment) and visualized using seaborn. The results verified trends in effects of injury and recovery after treatment on extent and solidity. Both findings support the expected shift from a circular shape of microglia in the injured state to more branched in the healthier state. The Otsu thresholding is limited in its accuracy, and, hence, these results provide an opportunity to optimize cell segmentation protocol for higher quality thresholded images. The results of this work have the potential to be applied to various forms of injury and cell types.

Single amino acid conformational preferences on a substrate are invaluable to our understanding of how conformational propensities are dictated by a peptide’s sequence. The efficiency of understanding SAP/SLAM (self-assembling peptides on single layer atomic materials) interfaces is of high importance. An efficient way of performing computational modeling of a peptide’s free energy landscapes is needed to predict the folded structures on solid surfaces towards designing bio/nano interfaces, a key to bioelectronics and biosensor developments. Graphene, single-atomic layer graphite, is an ideal substrate for a dodecapeptide to bind and spontaneously self-organize to form ordered biomolecular structures on the surface. For the purpose of shortening computation times of peptides at graphene interfaces, sampling each amino acid’s free energy landscape in terms of the peptide’s natural torsional configuration, such as the phi and psi angles of the peptide backbone on graphene is studied. This is due to the fact that a predetermined natural starting point, such as the peptide’s lowest energy structure will allow for quicker convergence of the system and more accurate structure prediction. In an effort to retrieve this data, here we simulate each amino acid using the enhanced sampling computational technique Metadynamics. Using the computational modeling and random sampling of the lowest energy wells, we aim to aid in the determination of low energy preferences on conformational landscapes on graphene towards more predictive design of soft bio/nano interfaces for practical implementations. 

Since its discovery, graphene has become a prominent material of interest for advanced bioelectronic and biomedical applications such as diagnostics, drug delivery, and imaging. This two-dimensional atomically thin sheet of sp2 hybridized carbon has exceptional electronic properties and is well suited for the development of highly sensitive and selective biosensors when paired with biomolecular adlayers. Additionally, by controlling the biomolecular orientation, conformation, and assembly structure of the adlayer, device functionality and performance can be fine-tuned. In this work, we focus on the conformational properties of three graphene-binding peptides, GrBP5-WT, GrBP5-M2 and Truncated GrBP5-M2, that form strongly adhered self-assembled adlayers at graphene surfaces. While these peptides are chemically very similar, experimental observations revealed they demonstrate opposite assembly phenomena upon thermal stimulus. Herein, enhanced sampling molecular dynamics simulations were employed using GROMACS simulation package with the PLUMED plugin to unravel how specific peptide conformations lead to the unexpected assembly behavior. Self-assembling peptide conformations were identified by comparing their computationally derived binding energies to experimental energetic data obtained from a scanning probe microscopy based molecular energetic analysis. The self-assembly structure of peptide on the graphene surface could form a module on the biosensor and used to detect specific proteins or peptides. With a better understanding of the peptide self-assembly mechanism, it would be significant to the development of biosensor.

Hydrogels are a versatile biomaterial commonly used for tissue engineering and drug delivery. In particular, photodegradable hydrogels have facilitated research breakthroughs in multiple fields, including organ development, disease progression, and blood vessel formation. While these reports have contributed greatly to the literature using in vitro experiments, current photodegradable designs are unable to function inside the human body due to their insensitivity to low energy light. Nearly all photodegradable hydrogels incorporate ortho-nitrobenzyl moieties as photosensitizers, which responds to UV light, a wavelength that does not penetrate complex tissue. In order to expand the applications of these photodegradable hydrogels a new crosslinker is needed that cleaves in response to visible light. I am working to develop a new photodegradable crosslinker based on ruthenium polypyridyl linker complexes, which can be structurally tuned to respond to visible light irradiation, leading to exchange of a ligand with water and rapid hydrogel degradation. We have modified the complexes with a reactive azide handle for site-specific incorporation into hydrogel biomaterials that can be transplanted to or formulated within specific bodily locations. Our preliminary results suggest the Ruthenium polypyridyl complexes degrade in the visible spectrum, and previous experiments have concluded that click chemistry allows for PEG-hydrogel incorporation with azide-modified crosslinkers. In this poster, I will describe the synthesis and characterization of one model Ru-based linker, including its photolysis, stability, and applications of the complex in the development of dynamic biomaterials for drug delivery and cell growth in vivo.