Utilizing machine learning models to synthesize perturbation data to generate perturbation heatmap graphical user interfaces

US12374429B1 · kind B1 · utility

Assignee

• Recursion Pharmaceuticals, Inc. · US

Inventors

• Marta Marie Fay · Salt Lake City, US
• August ALLEN · Boulder, US
• Eugene Yin-Chung Ting · Toronto, CA
• Lina Maria Nilsson · Salt Lake City, US
• Condie Thomas Swallow, II · West Valley City, US
• Michael Haines · Hillsboro, US
• Denton Hallar GREENFIELD · Evansville, US
• Kristin Ann Clark · Lehi, US
• Lovina Roundy · Orem, US
• Michael Joseph Uloth · Dundas, CA
• Sara Marjean Moore · Boise, US
• Shweta Deepchand Bhandare · Boulder, US
• Ted Douglas Monchamp · Nashua, US
• Summer Walid Elias · Salt Lake City, US
• Berton Earnshaw · Cedar Hills, US
• Mason L. Victors · Salt Lake City, US
• Safiye Celik · Sudbury, US
• James Benjamin Taylor · Midlothian, US
• Andrew David Blevins · Redwood City, US
• James Douglas Jensen · Farmington, US
• Jacob Carter Cooper · Sandy, US
• Conor Austin Forsman Tillinghast · Salt Lake City, US
• Seyhmus Guler · Salt Lake City, US
• Kyle Rollins Hansen · Kaysville, US
• Sarah Jordan DeVore · Salt Lake City, US
• Tongzhou Shen · Surrey, CA

Key dates

• Filing date: Dec 1, 2023
• Grant date: Jul 29, 2025
• Expiry date: Dec 1, 2043

Classification

• Technology area (CPC G): Physics
• CPC primary: G06T2207/30072
• WIPO field: Computer technology
• WIPO sector: Electrical engineering

Abstract

The present disclosure relates to systems, non-transitory computer-readable media, and methods for embedding perturbation data via a machine learning model and filtering, aligning, and aggregating the embeddings to generate a genome-wide perturbation database for real-time generation of perturbation heatmaps. In particular, in one or more embodiments, the disclosed systems can receive a plurality of perturbation images portraying cells from a plurality of wells corresponding to a plurality of cell perturbations. Further, the systems can generate, utilizing a machine learning model, a plurality of well-level image embeddings from the plurality of perturbation images. Moreover, the systems can align, utilizing an alignment model, the plurality of well-level image embeddings to generate aligned well-level image embeddings. Additionally, the systems can aggregate, according to perturbations of one or more perturbation experiments, the well-level image embeddings to generate perturbation-level image embeddings. Furthermore, the systems can generate perturbation comparisons utilizing the perturbation-level image embeddings.