The AI guide

What it's good at

• Identifying the system: protein name, organism, function, related PDB entries.
• Explaining the representation: what cartoons, ribbons, surfaces show and when to switch.
• Pointing at chains, residues, and ligands by name when the topology has them labelled.
• Summarising the simulation conditions from the uploader's description (force field, length, temperature).
• Suggesting questions worth asking next based on the system at hand.

What it isn't

The guide is not a structural biologist. It does not invent residue numbers or interactions that aren't in the topology, and it won't speculate about the underlying biology in ways that go beyond the cited sources. If you ask something it can't answer from the simulation and the linked databases, it'll say so rather than guess.

How it's grounded

Each answer is generated against three context sources, in order of priority:

1. The simulation's own topology and metadata (chain IDs, residue counts, uploader description).
2. UniProt and the Protein Data Bank, looked up by the related PDB IDs the uploader linked or that the topology mentions.
3. A small curated library of explanatory write-ups on canonical systems (lysozyme, T4 lysozyme variants, common kinase domains, ribozymes, etc.).

Citations appear inline at the end of each paragraph. Click one to jump to the source.

Asking good questions

The guide does best when you ask for a specific kind of help:

• "What am I looking at?" — orientation question.
• "Where is the active site?" — points the camera at a residue or region.
• "Explain the difference between cartoon and ribbon here." — pedagogical / representational.
• "What residues move the most over this trajectory?" — runs a quick RMSF query on the simulation.