Skip to content

CLAUDE.md

Latest commit

 

History

History
71 lines (52 loc) · 2.52 KB

CLAUDE.md

File metadata and controls

71 lines (52 loc) · 2.52 KB
Raw

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project Overview

Educational tutorial for glycoproteomic (proteomics with post-translational modifications) data analysis using R and Python. The project teaches bioinformatics workflows for processing, normalizing, analyzing, and visualizing mass spectrometry proteomics data.

Project Structure

  • Chapter 1: R basics and package introduction
  • Chapter 2: Data normalization (sample loading, TMM) with UMAP visualization
  • Chapter 3: Differential expression analysis using limma
  • Chapter 4: Gene ontology and pathway enrichment with clusterProfiler
  • Chapter 5: Structural analysis using localCIDER and StructureMap

Each chapter contains:

  • .Rmd files (R Markdown source - primary code)
  • .md files (rendered output for GitHub)
  • training_data/ folder with example datasets

Running Code

R Markdown files: Open in RStudio and knit, or use:

rmarkdown::render("Chapter_X/file.Rmd")

Python scripts: Called from R via reticulate:

library(reticulate)
use_condaenv(condaenv = '/opt/anaconda3/envs/structure_analysis', required = TRUE)
source_python("script.py")

Key R Packages

  • Data manipulation: tidyverse (dplyr, tidyr, purrr, stringr)
  • Data I/O: readxl, writexl, readr
  • Visualization: ggplot2, ggpubr, ComplexHeatmap
  • Statistics: rstatix, limma, edgeR
  • Bioinformatics: clusterProfiler, org.Hs.eg.db, AnnotationDbi
  • R-Python bridge: reticulate

Key Python Packages

  • Core: pandas, numpy, matplotlib, seaborn
  • Dimensionality reduction: umap
  • Structural analysis (Chapter 5): localcider, structuremap

Code Style

  • Uses tidyverse conventions with pipe operators (|>)
  • 2-space indentation (configured in .Rproj)
  • UTF-8 encoding
  • Data follows tidy data principles (each variable a column, each observation a row)

Data Flow Pattern

  1. Raw mass spectrometry results (CSV/TSV from search software)
  2. PSM filtering and quality control (XCorr, PPM thresholds)
  3. Protein-level aggregation
  4. Normalization (Sample Loading → TMM)
  5. Statistical analysis (limma for differential expression)
  6. Visualization and enrichment analysis

Important Notes

  • Large data files (.csv, .xlsx, .tsv) are gitignored - training data must be downloaded separately or generated
  • Python integration requires conda environment setup for Chapter 5
  • Column names in proteomics data often contain special characters - use name_repair = 'universal' when reading