This is the official Github repository for the project TEMPO: A Novel Hybrid Neuromorphic Framework for Efficient and Interpretable Integration of Multi-Omic High-Dimensional Oncology Data (PI: Lucia Nanda). TEMPO (TEmporal Multi-Omic Platform for Oncology) is a novel ANN-SNN hybrid neural network framework designed to facilitate efficient integration of complex oncology data. Rather than a traditional frontend-backend model, TEMPO implements a fully neuromorphic pipeline with embedded advanced ANN features such as encoders, routers, fusion layers, dissolving the boundaries between the ANN and SNN components. TEMPO also implements a variety of advanced features/techniques designed specifically to target challenges associated with the nature of cancer multiomics (high dimensionality, low sample sizes, missing data, class imbalance, etc), including hybrid context-aware encoding, semantic addressing, differentiable categorial routing, balanced focal learning, and weighted random sampling. More information about this software approach can be found in 'Introduction.ipynb'.

This project used publicly available deidentified data from The Cancer Genome Atlas (TCGA) GBM and PANCAN datasets, accessed from the Xena browser. To repeat developmental Phase 1 experiments in this repository, download: Methylation27k (n=288), Phenotypes (n=629), and IlluminaHiSeq (n=172), then run the preprocessing scripts outlined in 'Introduction.ipynb'. For final Phase 2 experiments, download: Batch effects normalized mRNA data (n=11,060), RPPA (n=7,744), gene-level copy number (gistic2) (n=10,845), Curated clinical data (n=12,591), and Molecular subtype (n=7,734), and run the corresponding preprocessing scripts. This project primarily used Python for machine learning programming and utilized Colab/Jupyter Notebooks.

Development files were developed privately since August 2025 before being uploaded and released publicly on GitHub. Currently, a complete TEMPO Python package is being developed that will include the model code, a wrapper, tests, and examples, allowing for accessible, clean, and convenient utilization of TEMPO.

Feel free to reach out to luciananda2010@gmail.com with any questions or comments.

Xena TCGA GBM: https://xenabrowser.net/datapages/?cohort=TCGA%20Glioblastoma%20(GBM)&removeHub=https%3A%2F%2Fxena.treehouse.gi.ucsc.edu%3A443

Xena TCGA PANCAN: https://xenabrowser.net/datapages/?cohort=TCGA%20Pan-Cancer%20(PANCAN)&removeHub=https%3A%2F%2Fxena.treehouse.gi.ucsc.edu%3A443