My research spans computational protein engineering, molecular virology, synthetic biology, global health equity, and evidence-based medicine. Every project is built on the same foundation: methodological rigor, reproducible analysis, and a question worth answering.
Leading a computational modeling project engineering novel mutations in human granzyme B (hGB) to simultaneously resist serpin B9 (PI9) inhibition — a tumor immune escape mechanism — while eliminating non-specific binding to non-target cells. The core challenge: combining the PI9-resistance mutation R201K with charge-reducing CS mutations (GrB^R201K-CS) abolishes enzymatic activity. Performing extended 500 ns molecular dynamics simulations on FZ Jülich's HPC infrastructure to elucidate allosteric communication pathways from CS-modified surface loops to the catalytic triad. Analyzing RMSF, PAD index, substrate contacts, and bond networks to pinpoint how the double mutant disrupts catalytic geometry. Designing 5–8 new in silico variants (alternative neutral substitutions, compensatory stabilizing mutations); top 2–3 candidates to be expressed, purified, and experimentally validated in Prof. Barth's UCT laboratory.
Working under Prof. Darren Martin, developer of the Recombination Detection Program (RDP), on molecular evolution, viral recombination, and the evolutionary dynamics of RNA viruses. Conducting computational analysis of viral sequence data to explore how recombination drives or limits viral fitness. Collaborating alongside PhD researchers including Ryan Hisner, applying Python, R, SQL, Pandas, and bioinformatics tools to address fundamental questions about mutation patterns and natural selection.
Mosquito-borne diseases cause 390 million infections annually, disproportionately affecting low- and middle-income regions. Engineered a Wolbachia-compatible expression plasmid producing dsRNA targeting NS5 polymerase mRNA, triggering Dicer-2-mediated cleavage of viral transcripts across Zika, Dengue, and Yellow Fever. Using S. cerevisiae (BY4741) as proof-of-concept host. NS5 coding sequences aligned with CLUSTAL and BLAST to identify conserved cross-species targets; antisense strand folding optimized with ViennaRNA. Experimental validation via RT-qPCR, northern blotting, western blotting, and mass spectrometry.
Audited all 98 primary human clinical studies cited in NCCN Kidney Cancer Guidelines v3.2022, building an original dataset of 44,636 participants across 46 countries. Applied PRISMA 2020, representation quotients (RQ), HHI, and weighted demographic pooling to analyze geographic distribution, demographic reporting completeness, and racial/ethnic representation gaps. Key findings: enrollment is concentrated in North America and Western Europe; Black participants are enrolled at roughly half their SEER-based incidence share; race is usable in only about one-fifth of studies; Hispanic ethnicity is rarely reported explicitly. Published through NSRI × MIT Critical Data Bias Journal.
Leading a systematic review investigating the relationship between community-level socioeconomic deprivation and histological features of liver disease. Incorporates Monte Carlo heterogeneity assessment to rigorously quantify variability across studies. Designing the review protocol and synthesizing evidence on how area-level deprivation correlates with liver histology findings.
Computationally designed and validated a 21-nucleotide siRNA targeting the conserved 5' UTR of Sindbis virus (SINV), a mosquito-borne alphavirus with no approved treatments. Used BLASTN against the human RefSeq RNA database and exhaustive seed-region k-mer searches (6–8 nucleotides, guide positions 2–9) to confirm no significant off-target matches to human transcripts, perfect conservation across 102 SINV isolates, and no cross-reactivity with Chikungunya. Contextualized findings within the biological reality that classical RNAi antiviral responses are weak in most mammalian somatic cells due to dominant interferon pathways—identifying CNS neurons as a relevant exception.
Co-designed a synthetic biology system using engineered E. coli swarms in a sprayable hydroxyethyl cellulose gel for rapid MRSA biofilm detection and disruption. The system uses quorum-sensing receptors to detect MRSA signals, triggering a colorimetric response via chromoprotein AmilCP (colorless → blue/purple), while DNase I and Dispersin B enzymatically degrade the biofilm matrix. A built-in toxin-antitoxin kill switch ensures biocontainment. Paired with a mobile app for contamination mapping and cleaning guidance. Designed for hospitals, schools, gyms, and other high-risk environments—20-minute detection, low-cost, portable.
Developed a biodegradable oral capsule to capture and remove microplastic fibers from the human digestive system. Tested five formulations of chitosan, lecithin, and citric acid. The 33/33/33 balanced blend achieved the highest average microfiber removal efficiency at 23.04%, demonstrating synergy between chitosan's binding, lecithin's emulsification, and citric acid's pH modulation. Placed 3rd in the Forsyth County Social Studies Fair.
Explored whether a gelatin film embedded with silver nanoparticles could protect potatoes from fungal infection via quantum confinement effects. Results were contrary to the hypothesis: experimental potatoes developed significant mold and fungal growth while controls primarily experienced drying, suggesting excess moisture in the coating overwhelmed the antifungal benefits. Generated important lessons about nanoparticle concentration, delivery method, and moisture control in agricultural antifungal applications.
Research paper investigating the detrimental long-term economic impact of hosting the Olympic Games. Analyzed historical data from Montreal 1976, Beijing 2008, and Paris 2024 (including the $1.5B Seine cleanup), documenting severe cost overruns, underused venue maintenance, and long-term taxpayer burden. Examined rare profitable exceptions (LA 1984) and critiqued the IOC bidding process. Placed 3rd in the Forsyth County Social Studies Fair.
Contributing to the MIT Critical Data Bias Engine's development and social media presence, supporting efforts to identify and address bias in health data and research. Also working on Sakshi, a program designed to detect logical fallacies and unsupported claims within research papers to strengthen the integrity of published science.
Coyote — API-powered book generator. Python, API Development, Wikipedia, Wolfram Alpha integration.
SwingSync (Mar–May 2024) — AI system that detects flaws in a golf swing and provides corrective feedback. Built in Python.
QuikSpace (Jun–Aug 2022) — App for storing and organizing information and events. Python.
Busare (Mar–Jun 2022) — Interactive food delivery simulation app built with Tkinter and Python.