ML-Directed Evolution Using Protein Language Models and CNN
Summary
Solugen, Inc. published patent application US20260099710A1 for a machine learning method used in protein engineering. The method generates fitness libraries using protein language model-guided site selection, trains convolutional neural networks to predict protein fitness, and optimizes sequences through phase transition-based algorithms with heating and cooling cycles.
What changed
Solugen, Inc. filed a patent application for a method combining protein language models (PLM) with convolutional neural networks (CNN) for directed evolution of proteins. The method identifies mutagenesis sites using PLM-guided site selection, trains a CNN with three-component sequence representation (latent embedding, probability matrix, and normalized zero-shot scores), and optimizes protein sequences using a phase transition algorithm with dynamic score matrix updates implementing heating and cooling cycles.
For biotechnology and pharmaceutical companies developing protein-based products, this patent represents potential prior art in machine learning-directed evolution. Companies using similar computational approaches for enzyme or protein optimization should review the claims for potential licensing implications or design-around considerations.
What to do next
- Monitor for updates
Archived snapshot
Apr 11, 2026GovPing captured this document from the original source. If the source has since changed or been removed, this is the text as it existed at that time.
MACHINE LEARNING FOR DIRECTED EVOLUTION
Application US20260099710A1 Kind: A1 Apr 09, 2026
Assignee
Solugen, Inc.
Inventors
Carlos Gomez Uribe, Japheth Gado
Abstract
A method for protein engineering includes generating a fitness library by performing protein language model (PLM)-guided site selection to identify favorable mutagenesis sites and creating protein variants, training a machine learning model to predict protein fitness from sequence data, wherein the machine learning model comprises a convolutional neural network (CNN); and optimizing protein sequences using a phase transition-based algorithm that dynamically updates a score matrix A of dimensions L×20 using cumulative statistics from sampled sequences and implements heating and cooling cycles to maintain system criticality. The CNN has a three-component sequence representation comprising a latent embedding matrix of shape L×1280, a probability matrix of shape L×20, and a feature vector containing 7 values including normalized zero-shot scores, where L represents the number of residues in the protein, wildtype subtraction normalization applied to the latent embedding matrix, dual parallel convolution processing paths, and percentile-based pooling.
CPC Classifications
G06N 3/08 G06N 3/0464
Filing Date
2025-10-03
Application No.
19349237
Related changes
Get daily alerts for USPTO Patent Applications - AI & Computing (G06N)
Daily digest delivered to your inbox.
Free. Unsubscribe anytime.
Source
About this page
Every important government, regulator, and court update from around the world. One place. Real-time. Free. Our mission
Source document text, dates, docket IDs, and authority are extracted directly from USPTO.
The summary, classification, recommended actions, deadlines, and penalty information are AI-generated from the original text and may contain errors. Always verify against the source document.
Classification
Who this affects
Taxonomy
Browse Categories
Get alerts for this source
We'll email you when USPTO Patent Applications - AI & Computing (G06N) publishes new changes.
Subscribed!
Optional. Filters your digest to exactly the updates that matter to you.