Biomarker Identification via Interaction Network Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for identifying molecular biomarkers face challenges such as low reproducibility and reliability, as they often fail to work across different datasets due to their assumption of functional independence, which is not reflective of the complex dysregulation of biological systems in diseases.
Innovation Solution
A computer-implemented method that involves data mining biomedical text and bioinformatic data to identify and rank candidate biomarkers using interaction network models, excluding correlation and expression relationships to focus on direct connections and applying algorithms like Dijkstra's shortest paths to build tissue-specific interaction networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional methods assume functional independence of molecular biomarkers, then individual biomarker identification is simplified, but reliability and reproducibility across different datasets deteriorates
Solution Approach 1:
The patent combines multiple biomarkers into an interaction network model that captures their functional relationships. Instead of evaluating biomarkers independently, the system integrates them into a unified network where nodes represent biomarkers and edges represent interactions, allowing simultaneous evaluation of multiple biomarkers while accounting for their interdependencies.
Solution Approach 2:
The patent introduces an interaction network model as an intermediary layer between raw biomarker data and disease prediction. This network model serves as a mediator that transforms individual biomarker measurements into integrated network-level features, capturing system-wide dysregulation patterns while filtering out dataset-specific noise.
2Productivity
If genome-wide screening with high-throughput techniques is performed, then the quantity of potential biomarkers increases, but the precision and reproducibility of identified biomarkers deteriorates
Solution Approach 1:
The patent extracts and removes correlation and expression relationships from the high-throughput screening data to focus on direct functional connections. By filtering out indirect associations and noise from genome-wide screening, the system isolates the most relevant direct interactions, improving precision while maintaining the ability to process large datasets.
Solution Approach 2:
The patent segments the complex genome-wide screening data into manageable interaction networks organized by disease pathways and biological processes. This segmentation allows the system to process high-throughput data in modular units, evaluating specific pathway-level interactions rather than attempting to analyze all possible biomarker combinations simultaneously.
3Adaptability or versatility
If interaction network models include all types of relationships (correlation, expression, direct connections), then the comprehensiveness of the model increases, but the ability to identify direct causal biomarkers deteriorates
Solution Approach 1:
The patent implements dynamic weighting of different relationship types within the interaction network model. Rather than treating all relationships equally, the system dynamically adjusts the influence of correlation, expression, and direct connection data based on their reliability and context-specific importance, allowing the model to adapt to different disease contexts while prioritizing direct causal relationships.
Data Source
AI summary
The present invention relates to systems and methods for identifying a biomarker from associative and knowledge based systems and processes. Particularly, aspects of the present invention are directed to a computer implemented method that includes data mining one or more public sources of biomedical text, scientific abstract, or bioinformatic data using queries to identify database terms associated with one or more predetermined terms, scoring association(s) between each of the identified database terms and the one or more predetermined terms, determining a subset b based on the score of the association(s), developing an interaction network model comprising the database terms in subset b, interactions, and additional database terms using a combination of algorithms in a predetermined order, and identifying candidate biomarkers from the interaction network model based on a ranking of the database terms in subset b and the additional database terms in the interaction network model.