Projects

Predicting chemotherapy-induced peripheral neuropathy (CIPN): CIPN is a prevalant, painful, and dose-limiting toxicity. Currently, there are no established cures for it due to a lack of complete understanding of its mechanism. Furthermore, different patients react differently to the treatment and there are no established predictors of CIPN. This project is in collaboration with clinicians at Riley Hospital, Indianapolis and Dr. Bruce Cooper, Bindley Bioscience Center, Purdue University. We analyzed metabolite profiles from the blood samples of pediatric patients who underwent chemotherapy treatment for leukemia. We then used support vector classifier along with recursive feature elimination to identify a small set of biomarker metabolites that can accurately predict CIPN susceptibility in these patients. We then built a model that can accurately predict CIPN susceptibility, using this small set of metabolites. Predicting CIPN susceptibility can help in preparing a personalized treatment for these patients and improve their quality of lives. I used CARET library in R for model building and feature selection. This work is published here.

Investigating mechanism of CIPN: CIPN occurs due to a myriad of pathophysiological factors such as alterations in expressions of voltage-gated ion channels, calcium signaling, blockage of axonal transport, etc. All these factors are interlinked in a complex fashion. In my work, I analyzed the role of different voltage-gated ion channels in inducing spontaneous firing: an indicator of peripheral neuropathy. For this, I analyzed a mathematical model representing dynamics of a pain-sensing neuron (of Hodgkin-Huxley type), and used bifurcation theory to find factors that can switch the system from a stable steady state (implying no peripheral neuropathy) to spontaneous firing or oscillatory state (implying peripheral neuropathy). I also investigated how the chemotherapy drug can affect these ion channels. This work is published here.

Pain sensation mutations: Due to specific genetic mutations, the pain sensation threshold of few individuals change. On one hand, some of them feel a chronic burning pain even without any external stimulus, on the other, some do not feel any pain even after being severely injured. I used bifurcation theory to explore how mutations in kinetics of sodium channels in a pain sensation neuron can alter the pain sensation threshold. I found specific parameters in the kinetics of these channels that can shift the bifurcation points of this system, and hence impact the pain sensation threshold. A published version of this work can be found here.

Bifurcation analysis of a pain sensing neuron model: I also performed a detailed bifurcation analysis of this model. This system displayed mixed-mode oscillations for specific parameter values. I found bifurcations explaining this mixed-mode oscillations regime. This system also displayed chaotic behavior for some parameter values. This work is published here.
Project 2, 3 and 4 are in collaboration with Dr.-Ing Achim Kienle and Dr. Dietrich Flockerzi at the Max Planck Institute for Dynamics of Complex Technical Systems.

Population balance modeling of leukemic lymphocytes: I employed population balance modeling to obtain time and age distributions of leukemic and normal cells, and tried to optimize infusion time of a chemotherapy drug in order to maximize killing of leukemic cancer cells and minimize killing of normal cells.

Undergraduate thesis: I investigated different functions to develop a moment closure function for the master equation of the TNF-a signaling network in order to estimate noise propagation.

Tools for predicting crystal morphology: As an undergraduate summer intern at Purdue University in 2012, I developed 4 software to predict crystal morphology from miller indices, tomographic indices and different saturation conditions. Find the links to the software in my Software page.