Classifier Training for Microwell Plate Artifact Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

High throughput screening in microwell plates faces challenges due to executional artifacts, which are errors resulting from the experiment process itself, making it difficult to draw valid conclusions from the dataset, as manual analysis is time-consuming, error-prone, and subjective, leading to overlooked or misinterpreted anomalies and inconsistent identification of trends.

Innovation Solution

A method is developed to train a classifier using spatial information from heat maps to automatically detect executional artifacts by computing feature vectors, applying wavelet transforms, and executing machine learning operations to generate a trained classifier that classifies microwell plates based on labels associated with executional artifacts, enabling consistent and objective analysis.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual analysis of heat maps is performed by reviewers, then executional artifacts can be identified and annotated, but the process is time-consuming and reviewers cannot scrutinize all heat maps within available time

Engineering Contradiction:

Improvedetection accuracy of executional artifactsVSAvoidanalysis speed and throughput

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent replaces the mechanical manual review process with an automated computer-based system that uses machine learning models to analyze heat maps. The system automatically detects executional artifacts, identifies anomalous patterns, and generates annotations without human intervention, thereby eliminating the time constraint while maintaining or improving detection accuracy through consistent algorithmic application.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service by allowing the heat map analysis process to execute autonomously without requiring reviewer input at each step. The automated pipeline processes heat maps, detects artifacts, and produces results independently, freeing reviewers from time-consuming manual scrutiny while ensuring comprehensive analysis of all heat maps.

Inventive Principle:

Principle #25Self-service

2Measurement precision

If manual review processes are used to identify executional artifacts, then artifacts can be detected, but the process is subjective and results vary from reviewer to reviewer

Engineering Contradiction:

Improvedetection capability of executional artifactsVSAvoidconsistency of artifact identification

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The patent replaces subjective human judgment with objective algorithmic analysis. The machine learning model applies consistent criteria to detect executional artifacts across all heat maps, eliminating inter-reviewer variability. The system produces reproducible results based on programmed detection rules rather than subjective interpretation, thereby improving reliability while maintaining detection capability.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If reviewers perform cursory analysis to complete within available time, then some heat maps are analyzed, but executional anomalies are overlooked or misinterpreted

Engineering Contradiction:

Improveanalysis throughputVSAvoiddetection accuracy of executional anomalies

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The automated system performs comprehensive analysis of all heat maps without the time constraints that limit manual review. The machine learning model systematically examines each heat map in detail, detecting subtle anomalies that cursory manual review would miss. This substitution enables both high throughput and high accuracy by automating the analysis process to handle large volumes of data with thorough scrutiny.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

4Measurement precision

If manual analysis is used, then executional artifacts can be identified, but trends in executional artifacts over time cannot be detected

Engineering Contradiction:

Improveidentification accuracy of executional artifactsVSAvoidloss of temporal trend information

Core Design Contradiction:

Measurement precisionVSLoss of information

Solution Approach 1:

The automated system performs multiple functions: it not only identifies executional artifacts in individual heat maps but also aggregates results across time to detect trends. The system can analyze temporal patterns in artifact occurrence, providing insights into changing experimental conditions or equipment degradation. This multi-functionality preserves temporal information while maintaining accurate artifact identification.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20220027795A1Techniques for training a classifier to detect executional artifacts in microwell plates

Publication Date: 2022.01.27 RECURSION PHARMACEUTICALS INC

AI summary

In various embodiments, a training application trains a classifier to detect executional artifacts in experiments involving microwell plates. The training application computes spatial information based on a heat map associated with a microwell plate. The training application then computes a set of features based on the spatial information. Subsequently, the training application executes one or more machine learning operations based, at least in part, on the set of features to generate a trained classifier. The trained classifier classifies sets of features associated with different microwell plates with respect to labels associated with executional artifacts. Advantageously, the trained classier can be used to accurately and consistently detect executional artifacts across different experiments and over time.