Machine Learning Compound Selection Pipeline

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional discovery pipelines rely on non-data-driven decision heuristics, limiting the ability to predict the efficacy, safety, and dosage of new compounds or biological actives when transitioning from lab to field, leading to inefficiencies and inaccuracies in compound selection.

Innovation Solution

A system that transforms lab and field data into a format suitable for supervised machine learning models to determine input variable importance, generate logical rules, and predict treatment effects, thereby optimizing the selection of compounds or biological actives using a combination of machine learning and statistical models.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional non-data-driven decision heuristics are used in discovery pipelines, then the process is simpler to operate, but the accuracy and reliability of compound selection deteriorates

Engineering Contradiction:

Improveaccuracy of compound selectionVSAvoidcomplexity of data processing system

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system segments the data processing workflow into distinct modules: data transformation module, machine learning model execution module, rule generation module, and statistical model execution module. Each module handles a specific aspect of the analysis, making the complex system manageable and maintainable while improving compound selection accuracy through systematic processing of lab and field data

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system introduces machine learning models and statistical models as intermediary components between raw data and final compound selection decisions. These models act as mediators that process and interpret complex relationships in the data, generating predictions and rules that guide compound selection with higher reliability than traditional heuristics

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If traditional discovery pipelines are used, then the process is faster to implement, but the productivity and efficiency of compound selection deteriorates

Engineering Contradiction:

Improveefficiency of compound selectionVSAvoidtime for data transformation and analysis

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary data transformation and formatting before analysis, converting lab and field data into a standardized first format suitable for machine learning models. This preliminary preparation enables more efficient processing during the actual compound selection process, improving productivity while the automated nature reduces manual time investment

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system replaces manual, non-data-driven decision-making with automated machine learning models and statistical models. This substitution of mechanical decision processes with computational models significantly improves efficiency and productivity, enabling faster and more accurate compound selection without proportionally increasing implementation time

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If more data is collected and processed through machine learning models, then the reliability of predictions improves, but the use of energy and computational resources increases

Engineering Contradiction:

Improvepredictive accuracyVSAvoidcomputational energy consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system dynamically adjusts data transformation parameters and model execution parameters based on the complexity of the data and the specific analytical requirements. By optimizing these parameters, the system achieves high predictive accuracy while minimizing unnecessary computational energy consumption, balancing reliability with resource efficiency

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS11373121B2Method to increase discovery pipeline hit rates and lab to field translation

Publication Date: 2022.06.28 SAS INSTITUTE INC

AI summary

The computing device transforms lab data and field data into a first format suitable for execution with a supervised machine learning model to determine an input variable importance for a first set of input variables in predicting a field outcome. Based on the determination, the computing device generates one or more logical rules of decision metrics, selects the one or more input variables that yields a higher input variable importance, and generates one or more pass-fail indicators. The computing device combines the one or more pass-fail indicators and generates one or more prediction factor rules. The computing device transforms the field data and the one or more prediction factor rules into a second format suitable for execution with a model to determine a treatment effect for the one or more prediction factor rules. The computing device selects the prediction factor rule that maximizes the treatment effect.