Nikassh Bala

AI Systems for Genetic Analysis, Space Life Sciences & Multiscale Systems Engineering

---

🔬 Executive Summary & Research Profile

I engineer computational biology systems, deep graph learning pipelines, and multiscale dynamical solvers to solve complex challenges in Space Life Sciences, Epidemic Modeling, and Environmental Systems Engineering.

My core focus spans:

1. Graph Neural Networks (GNNs) for Drug Discovery: Link prediction on heterogeneous knowledge graphs integrating transcriptomics (NASA GeneLab) with bioactivity databases (ChEMBL, CTD) to identify drug candidates for microgravity-induced physiological stress.
2. Multiscale Compartmental ODE Solvers: Coupling within-host intra-cellular viral kinetics with population-level macro-epidemiology to capture non-linear disease dynamics and behavioral feedbacks.
3. Reactive Transport & Water Quality Engineering: EPANET-MSX multi-species biofilm dynamics and Global Sensitivity Analysis (Sobol/eFAST) for opportunistic pathogen persistence.

---

🌟 Featured Research Repositories

Repository	Research Domain	Key Methodologies
gnn-drug-repurposing-spaceflight-atrophy	Space Life Sciences & Bio-AI	Heterogeneous GraphSAGE, NASA GeneLab DEG Consensus, Ensembl Orthologs
Coupled-Epidemic-ODE-Solver	Multiscale Mathematical Biology	Stiff ODEs (Radau), Next-Gen Matrix $R_0$, 134-State SEPIAQRV
legionella-biofilm-msx-model	Environmental Systems Engineering	WNTR / EPANET-MSX, Biofilm Detachment, Sobol Global Sensitivity

---

🧮 Theoretical & Mathematical Foundations

1. Heterogeneous Graph Convolution (Spaceflight Therapeutic Discovery)

Node representation updates across relation types $r \in \mathcal{R}$ connecting Gene, Drug, and Ortholog entities:

$$\mathbf{h}_{v}^{(k)} = \sigma \left( \mathbf{W}_{\text{self}}^{(k)} \mathbf{h}_{v}^{(k-1)} + \sum_{r \in \mathcal{R}} \sum_{u \in \mathcal{N}_r(v)} \frac{1}{|\mathcal{N}_r(v)|} \mathbf{W}_{r}^{(k)} \mathbf{h}_{u}^{(k-1)} \right)$$

2. Intra-Host to Macro-Population Transmission Coupling

Coupling within-host viral load $V(\tau)$ to the force of infection $\lambda(t)$ via Hill kinetics:

$$\beta(V) = \beta_{\text{max}} \cdot \frac{V^n}{\text{EC}_{50}^n + V^n}, \quad \lambda(t) = \beta(V) \cdot \frac{\sum (I_{\text{symptom}} + \epsilon A)}{N}$$

---

🛠️ Technical Capabilities

• Deep Learning & Graph Mining: PyTorch, PyTorch Geometric (PyG), HeteroData, NetworkX, DeepSNAP
• Scientific Computing & ODEs: SciPy (solve_ivp), NumPy, SymPy, Matplotlib, Seaborn
• Bioinformatics & Systems Biology: DESeq2 / Limma differential expression, Ensembl REST API, ChEMBL API, CTD API
• Environmental Hydraulics: WNTR, EPANET-MSX, SALib (Sobol / eFAST Global Sensitivity Analysis)
• Software Standards: Modular Python packaging (src/), PyTest, Docker, GitHub Actions (CI/CD), LaTeX, BibTeX

---

📜 Citation & Academic Metadata

If you reference or build upon any of these computational frameworks in your research, please use the citation metadata provided in each repository's CITATION.cff.

@misc{bala2026computational,
  author       = {Bala, Nikassh},
  title        = {Computational Biology, Space Life Sciences, and Multiscale Dynamical Solvers},
  publisher    = {GitHub},
  year         = {2026},
  url          = {https://github.com/Nikassh}
}