Network Entropy Detection of Pre-Disease State Transitions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting a pre-disease state in complex diseases face challenges due to little apparent change before the tipping point, unreliable disease models, and limited individual-based prediction, especially with high-throughput data noise and computational complexity.
Innovation Solution
A detection device and method utilizing local network entropy based on statistical mechanics to identify a precursor to state transition by calculating microscopic entropy between factors and their neighbors, reducing noise interference and computation load, and selecting biomarkers for early disease detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional biomarkers and snapshot static measurements are used for detection, then the measurement process is simple, but the detection precision is insufficient to distinguish pre-disease state from normal state
Solution Approach 1:
The patent transitions from static biomarker measurements to dynamic time-series measurements, capturing the temporal evolution of multiple factors. This parameter change enables detection of the pre-disease state through analysis of dynamical patterns rather than single-point snapshots, improving detection precision while requiring sophisticated analytical methods.
Solution Approach 2:
The patent introduces the time dimension by collecting multiple measurements over time, transforming static detection into dynamic detection. This additional temporal dimension allows differentiation between normal and pre-disease states through analysis of how factors evolve, rather than relying solely on single-timepoint values.
2Reliability
If model-based prediction methods are used, then the detection framework is established, but the reliability is low due to individual variations in deterioration processes
Solution Approach 1:
The patent employs unsupervised learning methods that allow the system to automatically discover patterns and structures in the data without relying on pre-defined disease models. The algorithm self-organizes to identify the pre-disease state based on temporal dynamics, adapting to individual variations without requiring model reconfiguration for each patient.
Solution Approach 2:
Instead of forcing data to fit predetermined disease models, the patent inverts the approach by using data-driven methods to discover the underlying patterns. The system learns the structure of deterioration processes from the data itself, rather than imposing external model assumptions, thereby achieving better reliability across diverse individuals.
3Measurement precision
If individual-based prediction is implemented, then the detection accuracy for specific patients is improved, but the productivity is reduced due to limited samples available for each individual
Solution Approach 1:
The patent performs preliminary analysis by pre-processing and normalizing time-series data, extracting relevant temporal features before applying detection algorithms. This preliminary action prepares the data in a form that maximizes information content from limited samples, improving individual detection accuracy without requiring extensive additional data collection.
Solution Approach 2:
The patent transforms the detection problem by changing from analyzing raw data directly to analyzing derived temporal patterns and dynamical features. This parameter transformation allows the system to extract meaningful signals from limited individual samples, achieving high detection accuracy while maintaining efficiency.
4Measurement precision
If high-throughput data analysis is performed to detect DNB, then the detection capability is enhanced, but the loss of time increases due to huge computational requirements
Solution Approach 1:
The patent extracts and focuses on specific temporal patterns and dynamical features that are most indicative of the pre-disease state, rather than analyzing all possible data aspects. By extracting the most relevant temporal characteristics, the system achieves high detection capability while reducing the computational burden of analyzing the complete high-dimensional dataset.
Data Source
AI summary
The invention provides a detection device, method, and program capable of highly accurately detecting a pre-disease state that indicates a precursor to a state transition from a healthy state to a disease state. The following processes are carried out: a process of obtaining measured data on genes, proteins, etc. related to a biological object as high-throughput data (s1), a process of selecting differential biological molecules (s2), a process of calculating the SNE of a local network (s3), a process of selecting a biomarker candidate (s4), a process of calculating an average SNE across the entire network (s5), and a process of determining and detecting whether or not the system is in a pre-disease state (s6).