January 2024 – September 2024
In the Systems Biology Lab's past work, an approach that integrates genome-scale metabolic models (GEMs) into machine learning model development was developed in order to determine effective drug-combination therapies for those suffering from Mycobacterium tuberculosis (TB) infection.
In the previous iteration of this approach (CARAMeL), the GEM used was produced from M. tuberculosis transcriptomics data. While using this GEM proved effective at predicting drug combination outcomes when compared to literature, a better model could be developed using sMtb-RECON, a GEM which includes host-pathogen interactions. Thus, this project revolves around reworking and developing CARAMeL to work with this GEM in order to more precisely predict drug-combinatorics to combat M. tb infections in humans.
Relevant skills: MATLAB · Linux · Bash Scripting · Computing Cluster Usage · Next Generation Sequencing (NGS) Data · scRNAseq Data Handling · xOmics · Machine Learning (ML) · ML Refinement · Code Troubleshooting
Learn more about CARAMeL and the Systems Biology Lab at the University of Michigan:
May 2023 – December 2023
In patients with epilepsy or brain tumors, intracranial electroencephalography (iEEG) may be used as an accurate tool to read the activity of the cerebral cortex. The signals generated are cyclical in nature, taking on characteristic shapes which are visually identifiable.
Most analyses include Fourier approaches, which assume that these signals are sinusoidal in nature. However, iEEG signals tend to be non-sinusoidal, so these approaches tend to overlook the characteristic shapes of these signals. This project seeks to find methods to identify the latent shapes in iEEG data to improve and parameterize the information that can be extracted from these signals using signal processing and machine learning approaches in order to improve our knowledge about iEEG signals in patients with epilepsy or brain tumors, and how they compare to healthy patients.
Relevant skills: Python · MATLAB · PyTorch · Signal Processing · Machine Learning (ML) · Discrete Signal Algorithms · Code Troubleshooting · ML Refinement
Learn more about the Multisensory Perception Lab at the University of Michigan:
January 2023 – May 2023
Currently, hospital staff use lateral patient transport devices to move bariatric patients from their beds to the operating table, reducing strain on healthcare workers. These mat-like devices minimize friction under the patient but cannot be removed from under them on the operating table due to space constraints. Thus, at angles such as the Trendelenburg position (30-45° from horizontal), reduced friction can cause patients to slip, potentially leading to nerve damage from prolonged pressure or, in extreme cases, the patient slipping off the table.
To mitigate these risks and facilitate patient maneuverability before and after surgery, myself and a small team designed a specialized clamp used to secure the low-friction mat and prevent movement during surgical procedures.
Relevant skills: COMSOL · CAD (SOLIDWORKS) · Validation and Verification · User Needs Identification · ISO 9001/13485 · Project Collaboration · FMEA/Risk Analysis · FDA Compliance
The final design review, poster and video for this project are available below!
August 2022 – December 2022
Aseptic loosening is a fairly common long-term downside of total hip implant operations in which the prosthetic begins to loosen from the surgical insertion site seemingly without mechanical failure or infection.
For this project, myself and a small team addressed this problem by designing a drug-eluting hip implant, whose function is to release a growth factor that leads to increased growth and fixation of bone around the implant. This, in turn, allows for a better-fitting hip implant that is less likely to experience aseptic loosening.
Relevant skills: COMSOL · CAD (SOLIDWORKS) · FDA Compliance · Stakeholder and Market Analysis · Project Coordination · Engineering Design Process (EDP)
The design history file (DHF) for this project is available below!
January 2022 – May 2022
People with constant computer mouse usage often develop Carpal Tunnel Syndrome (CTS), a neuropathic disorder which causes pain and numbness in the median nerve at the carpal ligament. To address instances of CTS in people who require constant computer mouse usage, myself and a small team developed a mouse-less cursor control system to reduce reliance on currently used computer mouses.
This system utilizes electromyography (EMG) signals obtained from muscles of both wrists and calves for right and left click, and scrolling up and down, respectively. This intuitive design enhances user convenience by utilizing easily controlled muscle groups for cursor control, and provides validation for the realization for an alternative method of cursor control which prevents the development of CTS.
Relevant skills: LabVIEW · Python · Bioinstrumentation · EMG Signal Processing · Basic Circuitry · Project Coordination
The paper and video for this project are available below!