A method for metabolomics data annotation based on structure learning and spectral graph recursive algorithm

Abstract

The application discloses a metabolomics data annotation method based on structure learning and spectral graph recursion algorithm, and belongs to the field of metabolomics data analysis. The method is developed in view of the existing bottleneck of omics data analysis, and combines modeling of metabolic reaction rules, joint processing of omics data and intelligent annotation of metabolic networks. By introducing a structure feature learning model and a spectral graph recursion algorithm, multi-level mass spectrometry information is fused to realize high-confidence structure annotation of unknown metabolites. Meanwhile, the method supports automatic preprocessing, annotation and atlas construction of various mainstream mass spectrometry data formats, and has good data compatibility and algorithm expansibility. The application can significantly improve the coverage and accuracy of metabolite annotation, and is suitable for high-throughput metabolomics analysis scenes of multiple types of biological samples such as blood, urine and tissue, and provides efficient and reliable technical support for disease mechanism research, drug metabolism evaluation and biomarker screening and the like.

Images