pkaScope

pkaScope is a Streamlit web app for exploring protein protonation states as a function of pH. It runs PROPka through its Python API, computes continuous protonation fractions for titratable sites, and exposes the result through an interactive structure viewer and residue tables.

Live demo: Hugging Face Space

Current Features

• Load an example PDB from examples/, upload a local .pdb file, or fetch a structure by RCSB PDB ID.
• Run PROPka directly in Python and cache the parsed result in Streamlit.
• Sweep pH with a slider from 0.0 to 14.0.
• Inspect per-site pKa, protonated fraction, current charge, and a simple state label.
• View folded and unfolded charge profiles across pH.
• Run pdb2pqr hydrogen-bond placement/optimization from a dedicated pane.
• Compare before/after structures and download optimized .pdb and .pqr outputs.
• Inspect raw PROPka summary and determinant output in a dedicated pane.

Running Locally

This repository uses uv for Python workflows.

git clone --recurse-submodules git@github.com:giorginolab/pkaScope.git
cd pkaScope
uv sync
uv run streamlit run main.py

If you already cloned the repository without submodules, run git submodule update --init --recursive once before uv sync.

Then open the local Streamlit URL shown in the terminal.

H-Bond Optimization Pane

The H-Bond Optimize tab runs pdb2pqr on demand when you click Run pdb2pqr optimization. It uses the same global pH slider as the other panes and caches successful results per structure and pH.

After completion, the pane shows:

• run summary metrics
• before/after Mol* previews
• download buttons for optimized PDB and PQR files
• a compact diagnostics section

How the First Viewer Version Works

The current viewer uses a modified streamlit-molstar to render the structure. That component does not yet expose a convenient Python-side API for per-residue coloring or selection, so the app currently rewrites the loaded PDB before display:

• occupancy stores the average protonated fraction for each titratable residue
• B-factor stores transition intensity scaled to 0-100

This keeps the pH-dependent state available in the structure artifact now, while leaving room for richer visual styling later.