How pivotbio/proven40-moa-paper-code works
Project type and purpose
• This is not a typical client‐server or microservice system but a reproducible data-analysis pipeline for a scientific paper. Its goal is to take raw experimental datasets (lab assays, field trials, on-farm trials) and produce figures and tables that support the study’s findings.
File structure
• Top-level files:
• environment.yml – conda specification for Python/R dependencies
• load_environment.R, packages_citations.txt, session_info.txt – R environment bootstrap and reporting
• README.md, LICENSE – documentation and licensing
• Three main directories, each one a self-contained analysis module:
• 1_LAB_ASSAYS
• 2_FIELD_TRIALS
• 3_ONFARM_TRIALS
• Each module contains:
• datasets/ – raw or pre-processed input data (Excel, CSV, RDS)
• *.ipynb – Jupyter notebooks with code cells, sequentially numbered
• *.html – exported notebooks for preview
• figures/ – serialized outputs (e.g., .rds objects or image files)
Insights from the README
• The three directories correspond to the three experimental tiers reported in the paper.
• Each analysis is intended to be run in order, within a conda environment named “moa_paper” that combines Python and R kernels.
• Notebooks produce both HTML previews and underlying figures/tables used for the manuscript.
a. Main components
• Processing environment: Conda environment (Python + R) / Jupyter Server
• Analysis notebooks: Three sequential modules of notebooks
• Outputs: Figures and tables stored in each module’s figures folder + HTML previews
• Version control: Git repo with README, environment spec, session info
b. Relationships & data flow
c. Architectural patterns & principles
• Modular separation by experiment type (each module independent)
• Reproducibility via environment specification (environment.yml)
• Sequential, linear data-to-result workflow per module
• Use of literate programming (Jupyter notebooks combining code, narrative, and results)
d. Key technologies & libraries
• Anaconda/conda
• Jupyter notebooks (.ipynb) with Python and R kernels
• R (load_environment.R, .rds outputs)
• Pandas (CSV/Excel ingestion), ggplot2 or matplotlib for plotting
• Git for version control