Teaching reaction informatics via formal graph theory¶
SynEdu is a sequence of nine executable notebooks on molecular graphs, atom mapping, reaction rules, and one-step prediction.
Start with S01 → Browse all talktorials Set up locally
Course structure¶
The lessons are grouped into three stages. They can be read in order or used individually.
S01–S03 · Molecules to graphs
Represent molecules explicitly, match graph structures, and align them with maximum common substructures.
S04–S07 · Graphs to reaction rules
Connect atom maps to ITS graphs, DPO rewriting, canonicalization, and reusable rule libraries.
S08–S09 · Rules to products
Apply rules forward and backward, evaluate candidates, and tune context specificity.
Lessons¶
Each notebook introduces one part of the reaction-informatics workflow.
Molecules to labeled graphs
Graph homomorphisms
Maximum common substructures
Atom mapping and ITS graphs
Reaction rules as graph rewriting
Deterministic reactions and rules
Reusable DPO rule libraries
One-step reaction prediction
Context radius and specificity
Run and reuse¶
Use JupyterLab locally or launch a portable lesson in Colab or Binder.
The Python API contains the graph and reaction helpers used in the lessons.
Fix a lesson, report an issue, or add an example.
SynEdu uses RDKit, NetworkX, and MyST. Related projects are listed on the SynEco page.