Systems and methods for effect size optimization using cell annotations
A classification model with barcode annotations and a Z' based cost function optimizes effect size in fluorescence microscopy bioassays, addressing inefficiencies in feature selection and metric parameter tuning, thereby accelerating molecular biology research.
WO2026156305A1PCT designated stage Publication Date: 2026-07-23ARACELI BIOSCIENCES INC
View PDF 0 Cites 0 Cited by
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- ARACELI BIOSCIENCES INC
- Filing Date
- 2026-01-16
- Publication Date
- 2026-07-23
AI Technical Summary
Technical Problem
The process of selecting features and honing metric parameters for fluorescence microscopy bioassays is time-consuming and iterative, especially when dealing with new treatment and effect combinations, leading to inefficiencies in molecular biology research.
Method used
A method using a classification model trained with barcode annotations and a Z' based cost function to automatically optimize effect size by classifying cells, reducing the need for manual labor and trial-and-error in selecting and weighting measurements.
Benefits of technology
This approach accelerates molecular cell biology research by providing a rapid and efficient method to determine features that maximize Z', thereby improving bioassays and reducing development time and costs.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure US2026011667_23072026_PF_FP_ABST
Abstract
Methods and systems are provided herein for automatic effect size optimization for separating classes of cells defined by barcode annotations. In an example, a method includes obtaining a field of view (FOV) image of a sample including a plurality of cells, determining an effect size of a perturbation to a first portion of the plurality of cells in the FOV image using a classification model, the classification model trained based on training data including a plurality of first instance images of cells of a first class and a plurality of second instance images of cells of a second class and further based on an effect size loss function, wherein the cells of the first class are identified based on an annotation exhibited by each cell of the first class, and outputting an indication of the effect size.
Need to check novelty before this filing date? Find Prior Art