A unified ecosystem for macromolecular design tools and automated workflows.
The Macromolecular Nexus (MacromNex) is a research-driven GitHub organization focused on the intersection of Geometric Deep Learning, Molecular Physics, and Synthetic Biology. Our mission is to maintain a unified ecosystem for State-of-the-Art (SOTA) methods while pioneering novel solutions for the most challenging problems in macromolecular design.
Comprehensive design platform for proteins, enzymes, and antibodies.
- Fitness Modeling: Modeling protein property given the experimental data, e.g. activity, stability, specificity, etc.
- Peptide Binders: Designing high-affinity binders for protein-protein interactions (PPIs).
- Antibody Engineering: Deep learning pipelines for CDR generation, paratope-epitope matching, and affinity maturation.
-
Enzyme Design: Optimization for catalytic turnover (
$k_{cat}$ ), thermostability, and regioselectivity.
Specialized research on constrained macrocycles and cyclic peptides.
- Structure Prediction: Predicting the (complex) structure of cyclic peptide binders.
- Cyclic Peptide Binder Design: Designing cyclic peptide binders given a protein target.
- Multi-objective Optimization: Optimization multiple properties of cyclic peptides using generative AI and reinforcement learning.
Advanced tools for designing functional and structural nucleic acids.
- RNA Design: 3D structure-to-sequence design for functional RNAs and aptamers.
- DNA Nanotechnology: Computational tools for structural DNA design and synthetic promoter engineering.
- Complexes: Modeling and designing Protein-RNA/DNA interfaces.
We are an open-science initiative. We welcome contributions from:
- Computational Biologists (Benchmarking and domain expertise)
- ML Engineers (Model optimization and architecture)
- Experimentalists (Data sharing and wet-lab validation)
Contact: [charlesxu90@gmail.com]
“Designing the future building blocks of life with artificial intelligence (AI).”