Biomarker Discovery via Dynamic Data Integration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for analyzing healthcare data predetermine variables, limiting the discovery of new or unknown relationships and biomarkers, which restricts the identification of key drivers of patient response to therapies.
Innovation Solution
A method that integrates molecular profile data and clinical records data to identify potential biomarkers for clinical outcomes by processing and analyzing data using statistical, machine learning, and artificial intelligence methods, generating causal relationship networks, and employing machine learning to select relevant biomarkers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional methods predetermine variables for analysis, then the analysis process is simplified and faster, but the ability to discover new or unknown relationships and biomarkers is limited
Solution Approach 1:
The patent implements a dynamic variable selection approach where the system automatically identifies and adapts relevant variables from large datasets during analysis, rather than relying on static preselected variables. This allows the analysis to evolve and discover novel relationships while maintaining efficient processing through automated relevance assessment.
Solution Approach 2:
The system changes the parameters of analysis by transforming predefined variable-based analysis into data-driven variable identification. It uses statistical methods and machine learning to dynamically adjust which variables are analyzed based on their relevance to clinical outcomes, thereby discovering new relationships without sacrificing analysis efficiency.
2Device complexity
If preselected variables are used for analysis, then the complexity of data processing is reduced, but the ability to identify key drivers of patient response is restricted
Solution Approach 1:
The patent replaces manual variable selection (mechanical approach) with automated machine learning algorithms that systematically identify key drivers. This substitution maintains manageable complexity through algorithmic automation while significantly improving the precision of identifying key drivers of patient response through data-driven insights.
Solution Approach 2:
The system introduces machine learning models as intermediaries between raw data and clinical insights. These models process large datasets, identify patterns, and select relevant variables automatically, thereby reducing processing complexity while enhancing the precision of key driver identification through sophisticated pattern recognition.
3Quantity of substance
If large amounts of medical data are collected from clinical trials, then the potential to identify biomarkers increases, but the challenge of analyzing this data to identify key drivers becomes greater
Solution Approach 1:
The patent segments the large medical dataset into manageable components using machine learning techniques. It divides the analysis into stages: data preprocessing, feature selection, model training, and validation. This segmentation reduces analysis complexity by breaking down the overwhelming task into systematic, automated steps that can be processed efficiently.
Solution Approach 2:
The system creates computational models that replicate and simulate the relationships in the medical data. By generating virtual representations of patient responses and biomarker relationships, it reduces the complexity of analyzing raw data directly, allowing researchers to study key drivers through simplified model simulations while preserving the essential patterns from the large dataset.
Data Source
AI summary
Disclosed herein are methods and systems for identifying one or more potential biomarkers for a clinical outcome related to administration of an agent. The method includes processing molecular profile data for a plurality of subjects where the molecular profile data includes data obtained before, during and/or after administration of an agent to the plurality of subjects. The method also includes processing clinical records data for the subjects, where the clinical records data includes clinical outcome data, integrating the processed molecular profile data and the processed clinical records data for the subjects and storing in a database as merged data, selecting two or more subsets of the merged data using one or more criteria based on the clinical records data to generate two or more selected data sets, and analyzing one or more of the selected data sets to identify one or more potential biomarkers for a clinical outcome related to administration of the agent.