Microbial Feature Selection for Machine Learning Constipation Diagnosis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for diagnosing constipation using machine learning models face challenges due to unprocessed samples causing noise and significant degradation in performance, making it difficult to accurately identify causative agents of constipation.

Innovation Solution

A method and apparatus that analyze a mixture of gut-derived substances and a gut environment-like composition to extract microbial data, select microbe-related features using a feature selection algorithm, and train a machine learning model to diagnose constipation, focusing on families like Lachnospiraceae, Erysipelatoclostridiaceae, and others.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If bacterial metagenome analysis is performed without sample culturing, then the analysis process is simplified and faster, but the accuracy of identifying causative agents of constipation deteriorates due to large bias among samples

Engineering Contradiction:

Improveanalysis process timeVSAvoidaccuracy of identifying causative agents

Core Design Contradiction:

Loss of timeVSMeasurement precision

Solution Approach 1:

The patent applies preliminary action by performing sample culturing before metagenome analysis. This preliminary culturing step enriches the bacterial samples and reduces bias among different subjects' samples, thereby improving the accuracy of identifying causative agents of constipation while maintaining a streamlined analysis process

Inventive Principle:

Principle #10Preliminary action

2Productivity

If machine learning model is trained using unprocessed samples, then the training process is simpler and faster, but the model performance deteriorates significantly due to large amount of noise in training data

Engineering Contradiction:

Improvetraining process efficiencyVSAvoidmodel performance

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent applies preliminary action by processing samples through metagenome analysis and feature extraction before training the machine learning model. This preliminary processing removes noise and extracts relevant microbial features, thereby improving model performance while keeping the training process efficient through automated pipelines

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent extracts relevant microbial features from the metagenome data, such as the abundance of specific bacterial genera (e.g., Lachnospiraceae, Erysipelatoclostridiaceae, Pseudomonadaceae, Prevotellaceae, Desulfovibrionaceae, Clostridiaceae, Gemellaceae, Bacteroidaceae, Streptococcaceae, Anaerofustaceae, Monoglobaceae, and RF39). This extraction process removes irrelevant noise from the training data, improving model reliability

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS12451254B2Method and diagnostic apparatus for determining constipation using machine learning model

Publication Date: 2025.10.21 HEM PHARM INC

AI summary

A method for determining whether constipation is present by using a machine learning model. The method includes a process of analyzing a mixture of a gut-derived substance collected from a subject and a gut environment-like composition, a process of extracting multiple microbial data based on an analysis result of the mixture, a process of selecting microbe-related features to be used in the machine learning model from the multiple microbial data based on a predetermined feature selection algorithm, a process of training the machine learning model with the microbe-related features, and a process of inputting, to the trained machine learning model, the microbial data collected from the subject to be tested and determining whether constipation is present.