Author: Julian Benson (B.S. Biochemistry & Molecular Biology) Target: Reactome / Open Genome Informatics (GSoC 2026)
Read the Full GSoC 2026 Architectural Proposal Here
This repository contains the data ingestion and optimization bridge designed to transition MP-BioPath from a literature-validation tool to a high-throughput predictive engine for genomic data.
The pipeline utilizes a polyglot architecture:
- Python/R (
src/module_a_ingestion): Processes raw RNA-seq differential expression data into the continuous 0.01 to 100 boundaries required by the MP-BioPath non-linear mathematical models. - Python (
src/module_b_graphs): Interfaces with the Reactome REST API to dynamically stitch sub-pathway logic graphs. - Julia/JuMP (
src/module_c_solver): Modifies the core objective function weights to resolve entity-set dilution and automate the 15% biological significance threshold.
/data: (Ignored via.gitignore) Local staging for MCB112 quantitative datasets and RNA-seq counts./src: Core module scripts for ingestion, API translation, and JuMP solver execution./docs: Project proposals, mathematical documentation, and architecture blueprints.
Wright, A.J., Orlic-Milacic, M., Rothfels, K. et al. Evaluating the predictive accuracy of curated biological pathways in a public knowledgebase. Database (2022). DOI: 10.1093/database/baac009