Continuous Variational Synthesis
q-bio.BM, cs.LG
Submitted: 2026-09-28
Updated: 2026-09-28
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
The gist: Biological machine learning was long bottlenecked by the ability to synthesize designed DNA.
Terminology
Abstract
Biological machine learning was long bottlenecked by the ability to synthesize designed DNA. Variational synthesis models control chemical reactions to physically manufacture quadrillions of designed sequences in DNA. However, training these generative models is challenging: constraints on chemical synthesis can force many parameters into a discrete space, limiting the ability to pre-train and fine-tune. In this article we train ``free'' variational synthesis models using stochastic gradient descent in continuous space, and then discretize with post-training quantization to impose hardware and wetware constraints. This enables variational synthesis models to satisfy stringent reward criteria, while still synthesizing diverse designs, achieving a strictly dominating quality-diversity Pareto frontier. We demonstrate by training variational synthesis models of enzymes, peptides, antibody CDRH3s, and regulatory DNA elements. In silico performance is maintained in vitro.
Related papers
- Speak to a Protein: An Interactive Multimodal Co-Scientist
- Learning Topological Representations of Protein Structure and Dynamics
- p2smi: A Python Toolkit for Peptide FASTA-to-SMILES Conversion and Molecular Property Analysis
- UNAAGI: Atom-Level Diffusion for Generating Non-Canonical Amino Acid Substitutions
- Co-folding with a Soup of Representations
- Non-Markovain Quantum State Diffusion for the Tunneling in SARS-COVID-19 virus