MBI Videos
2015 Undergraduate Capstone Conference
-
Corey WeistuchAbstract not submitted.
-
Dominic GrayAbstract not submitted. -
Austin FergusonAbstract not submitted -
Evan NorthCentral Pattern Generators (CPGs) are neural networks that are thought to control repetitive actions (chewing, breathing, walking, etc). While the existence of spinal CPGs in humans is still only a theory, it is a theory well supported by data. In studies of patients with knee injuries, a number of interesting patterns emerge in the periodicity of their leg movements. The main goal of this research is to create a simplified model that relies on mutually inhibitory neurons and sensory feedback from the system to show how a CPG could give rise to the patterns recorded in studies.
-
Sam HirniakAn understanding of cellular metabolism is crucial to a wide range of biotechnology and health-related activity, such as biofuel production and drug design. Unfortunately, the large scale of whole-genome metabolic networks means that some analytic tools, such as elementary flux mode analysis cannot be applied. Using a recently developed reduction method, we are able to reduce large metabolic models to more manageable, steady state equivalent networks. This reduction minimizes the number of elementary flux modes and hence facilitates flux mode analysis. In addition, we are addressing links between elementary flux mode analyses and signaling pathways. This involves generating a stoichiometric network with the same dynamics as a given signaling network, thus allowing computational tools such as elementary flux mode analysis to be applied in a new context.
-
Samantha FurmanAbstract not submitted. -
Lindsay MercerAbstract not submitted. -
Alex MercantiNo abstract submitted.
-
Joseph TharayilAbstract not submitted. -
Emily ShannonParkinson's Disease is a disorder of the nervous system, which causes a loss of dopaminergic neurons in the basal ganglia region of the brain. This loss reduces the quality of communication to neurons in the motor cortex, causing individuals to experience motor symptoms such as tremors, slowness of movement, and rigidity. One treatment option for this disorder is Deep Brain Stimulation (DBS), in which electrical impulses are applied to the subthalamic nucleus through implanted electrodes, connected by a wire lead to a pacemaker-like device in the chest. This stimulation affects neurons in the globus pallidus externa and substantia nigra pars reticula, improving the communication between neurons to relieve the associated motor symptoms. Due to the finite battery life of these devices, current research regarding DBS involves finding a way to minimize the power output of the device, while still retaining its ability to relieve symptoms. Data provided by Dr. Warren Grill's lab at Duke University was used in this project to explore how the information content of neurons changed for rats in a control case, with Parkinsonism, and undergoing DBS of different frequencies. Understanding how information changes under these different conditions, as well as under different frequencies of stimulation, can hopefully drive investigations into which frequency yields the best results.
-
Emily Nguyen, Amanda ReederWe use a spatial, stochastic, continuous time SIR model to investigate the dynamics of measles in a "country" of 25 cities. In this model, a person can be born, become infected, recover, immigrate to another city, become vaccinated, or die at any point in time. We are particularly interested in investigating how different vaccine regimes affect the disease's persistence and which, if any, lead to herd immunity. To that end, we study four different vaccination regimes and identify one that minimizes both the number of measles cases and the proportion of the population that must be vaccinated.
-
Alexander CaoStable heteroclinic channels (SHCs) offer an alternate modeling framework for rhythmic activity in biological motor control systems. They combine the robustness of central pattern generators (limit cycles) with the sensitivity of chain reflex systems. We introduce a two-dimensional noisy, twisted stable heteroclinic oscillator as an abstract model for neuronal motor control circuits. To investigate the interplay between stochastic and deterministic steering, we study the effects that both have on the mean return time of the orbits. Through direct numerical simulations, we show that the mean period has a non-monotonic response to noise intensity. We use a mean first-passage time (MFPT) approach to study this phenomenon more quantitatively. In the limit of small noise, semi-analytic results from the MFPT analysis agree with numerical results.
-
John CaughmanAbstract not submitted. -
Brandon Cannone, Rebecca ThiemTraditionally, behavioral and morphological data have been analyzed with Maximum Parsimony to estimate evolutionary histories by finding the fewest number of required changes in a tree connecting related species. In recent years, Bayesian estimation of molecular phylogenies has been heavily researched, but researchers typically assume a prior distribution that all phylogenies are equally likely. In this project, we integrate these schools of thought and data types by informing the prior distribution with the parsimony scores of the trees with respect to morphology and behavior. We implement this method on sample datasets of corbiculate bees and orb-weaver spiders.
-
Sharon FreshourAbstract not submitted.
-
Acara TurnerAbstract not submitted.
-
Taleah TyrellAbstract not submitted.
-
Alexander BeamsCould competition between antibiotic-vulnerable and antibiotic-resistant infections be used to manage resistance? Hypothetically, treating co-infected individuals those carrying both strains at once could promote antibiotic resistance. In the mathematical model, fitness tradeoffs between strains and different structures in the host population represent the effects of treatment schemes, with people divided into four compartments: susceptible, infected with one strain or the other, or infected with both simultaneously. We use the next generation matrix method to calculate R0 for the resistant and vulnerable strains, looking for situations where they experience competitive exclusion or coexistence, and focusing on those where resistance can invade. The model supports the hypothesis that treating doubly-infected individuals promotes antibiotic resistance, and in the future some simplifying assumptions shall be relaxed (perfect treatment implementation, for example) and in general more realistic treatment strategies that physicians use will be considered.
-
Terence SpeedI'm not sure when I did my first statistical consulting, but it must have been about 50 years ago. I have never really stopped doing it. I recently met with two people who are tracking immune cell fate determination, one cell at a time. They have been analyzing their data, and met several challenges. I have also been talking with someone analyzing the gut bacteria of mice, and separately, with a person who is studying the white blood cell response to stimulation by group A streptococcus. I have found over the years that consulting can morph into collaborative research, and that collaborative research can lead to statistical research. That's the way I like it.
My plan is to talk about some of the different aspects of statistical consulting and collaboration, and interdisciplinary research, focussing on what might be done on a university campus (rather than in business, industry or government). I'll speak about the pluses and the minuses, and I'll touch on the technical, psychological, social and political challenges that consulting, collaboration and interdisciplinary research throw up.