Automated Compound Selection via Historical Data Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for selecting compounds in drug discovery are inefficient and costly, relying on expert opinions and computational models that fail to accurately predict compound success, leading to high failure rates and financial losses.

Innovation Solution

A rigorous, automatic method for analyzing historical data to extract interpretable selection criteria and importance values for compounds, using a combination of top-down 'peeling' and bottom-up 'pasting' algorithms to identify criteria that distinguish successful compounds from unsuccessful ones, allowing for informed compound selection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If manual expert opinion-based criteria are used for compound selection, then decisions can be made with current available knowledge, but the approach cannot take into consideration large amounts of historical data and is limited by individual biases

Engineering Contradiction:
Improvehistorical data utilizationVSAvoiddata analysis complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent replaces manual expert analysis with an automated computational system that processes historical data. The system uses software algorithms to automatically analyze compound properties, assay results, and project outcomes, substituting human expert judgment with machine-based data processing to eliminate bias and handle large datasets efficiently.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent creates a virtual model of compound selection by copying and analyzing historical data patterns. The system learns from past successful and unsuccessful compounds by creating computational representations of their properties and outcomes, allowing the identification of selection criteria without manually reviewing each historical case.

Inventive Principle:
Principle #26Copying

2Measurement precision

If computational approaches such as QSAR models are used to predict compound properties, then predictions for individual properties can be made, but the models are insufficient to solve the challenges of predicting overall compound success and determining descriptor importance

Engineering Contradiction:
Improveprediction accuracyVSAvoiddescriptor importance information
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent implements feedback mechanisms where the system continuously learns from actual compound outcomes. By comparing predicted versus actual success rates and using techniques like cross-validation and bootstrapping, the system refines its models and identifies which descriptors most strongly correlate with successful compounds, creating a feedback loop that improves prediction accuracy over time.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent performs preliminary analysis of historical data to identify potential selection criteria before actual compound selection occurs. The system pre-processes historical outcomes and compound properties to establish weighted criteria and importance values that can be applied prospectively to new compounds, allowing predictions to be made based on pre-established patterns rather than ad-hoc analysis.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If over-aggressive filtering of drug pipelines is applied to eliminate unsuccessful compounds, then costs and failure rates may be reduced, but opportunities to find new therapies may be missed

Engineering Contradiction:
Improvecost efficiencyVSAvoidtherapy discovery success
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent dynamically adjusts selection criteria thresholds based on historical performance data and confidence levels. Rather than applying fixed aggressive filters, the system modifies criterion stringency based on the strength of evidence from historical data, allowing flexible filtering that adapts to different compound classes and project stages, thereby maintaining cost efficiency while preserving promising candidates.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2609209B1Compound selection in drug discovery
Publication Date: 2025.06.25 OPTIBRIUM
  • EP2609209B1 patent drawingFigure 1
  • EP2609209B1 patent drawingFigure 2A
  • EP2609209B1 patent drawingFigure 2B

AI summary

Methods and systems for determining the selection criteria that in its embodiments can distinguish compounds that successfully meet an objective from those that do not, determine the importance of selection criterion in selecting test compounds that have a high probability of achieving an objective and automatically apply the selection criteria to select test compounds with a high chance of meeting an objective.