Rapid identification of stress-tolerant microorganisms based on metagenomic data and transfer learning

CN122245440APending Publication Date: 2026-06-19INST OF SOIL SCI CHINESE ACAD OF SCI

Patent Information

Authority / Receiving Office: CN · China
Patent Type: Applications(China)
Current Assignee / Owner: INST OF SOIL SCI CHINESE ACAD OF SCI
Filing Date: 2026-03-23
Publication Date: 2026-06-19

AI Technical Summary

Technical Problem

Existing identification methods struggle to accurately capture potential associations between gene sequences when dealing with highly sparse and heterogeneous metagenomic data. Furthermore, traditional models are prone to overfitting and lack effective domain-adaptive strategies, resulting in low identification efficiency and high false alarm rates.

Method used

By constructing a multi-dimensional feature space, introducing a pre-trained deep neural network model and a transfer learning framework, and combining adversarial discrimination mechanism and attention enhancement strategy, feature distribution alignment and biological consistency verification are performed to screen out stress-resistant microorganisms with high confidence.

Benefits of technology

It significantly improves the efficiency and accuracy of identifying stress-resistant microorganisms, reduces the false alarm rate, and achieves rapid and accurate identification under limited sample conditions, with good scalability and adaptability.

✦ Generated by Eureka AI based on patent content.

Smart Images

Figure CN122245440A_ABST

Patent Text Reader

Abstract

This invention discloses a rapid identification method for stress-resistant microorganisms based on metagenomic data and transfer learning, belonging to the interdisciplinary fields of bioinformatics and artificial intelligence. During system operation, the target environment metagenomic data undergoes quality control and assembly, extracting multidimensional features such as amino acid composition, conserved domains, and collinear gene clusters. A pre-trained deep neural network is introduced to construct a transfer learning framework, and an adversarial discrimination mechanism is used to align the feature distributions of the source and target domains. An attention-enhanced classifier is combined to predict stress-resistant phenotypes, and biological consistency is verified through functional annotation and metabolic pathway enrichment analysis. This application, through the above technical solution, significantly improves the sensitivity and accuracy of identifying sparse stress-resistant signals, reduces the false alarm rate, and achieves efficient and automated discovery of stress-resistant microorganisms.

Need to check novelty before this filing date? Find Prior Art