CHEM 274B Micropresentation

MCODE-Based Protein Complex Detection (Vertex Weighting Project)

In this project, we use a Jupyter Notebook to implement the vertex-weighting stage of the MCODE algorithm to help identify potential protein complexes in a yeast kinase–substrate interaction network. Our goal is to turn the interaction data into a graph and then compute weights for each protein based on how dense its local neighborhood is.

The dataset we use comes from:

Bandyopadhyay S. et al., A human MAP kinase interactome (2010) https://doi.org/10.1038/ncomms1139

In our graph:

• Each protein is a node
• Each interaction between proteins is an edge
• We build the graph using NetworkX
• We compute vertex weights using:
  • k-core values
  • local density
• We visualize the weighted graph to highlight potential protein complexes

---

Files Included

• mcode.ipynb – Our main notebook with all code, explanations, and plots.
• Bandyopadhyay2010.xls – The protein interaction dataset (this must be in the same folder as the notebook for it to load correctly).

---

Requirements

To run everything, we need:

• Python 3.9+
• Jupyter Notebook or JupyterLab
• These Python packages:
  • numpy
  • polars==1.34.0
  • networkx
  • matplotlib
  • seaborn
  • plotly
  • jupyter

If we want to use a requirements.txt, it should look exactly like this:

numpy
polars==1.34.0
networkx
matplotlib
seaborn
plotly
jupyter

---

Contributors

This was a four-person project for CHEM 274B. The split below is taken from the team contribution document submitted with the assignment.

Jade Warren

• Base code for the MCODE notebook implementation and debugging
• Notes from papers for the slides

Carlos Lopez

• Algorithm literature search
• Notebook debugging
• 3D visualization

Manula Dombagahawatta

• Notebook debugging (primary)
• Original HIGH-PPI research

Brendan Dang

• Slide template and notes
• README
• Data and paper sources

Original repository: https://github.com/jadewithgreen/MCODE