Research And Implementation Of The Establishment Method Of Serum Standard Reference Material Database In Clinical Metabolome

In the post-genomic era,metabolomics is an emerging research field with unique development advantages.It has been widely used in clinical applications such as disease diagnosis,treatment monitoring,and disease rehabilitation.Correspondingly,various types of metabolite databases have been used for the large amount of omics data generated during the management research process effectively supports the further development of metabolomics.Although various types of metabolite databases have been established,the database of standard reference materials still cannot meet the actual needs.Standard reference materials have stable physicochemical properties and are often used in analytical method validation,process control,instrument calibration,etc.Therefore,in view of this situation,this thesis designs and implements a serum standard reference material database HRSMDB(Human Reference Serum Metabolome Database),which can be used in clinical metabolomics research,and further expands the functions of this database.The main contributions of the thesis:(1)Designed and implemented a serum standard reference material database,which provides serum standard reference samples prepared according to the standard production process,and the corresponding standard data of the samples;a large number of high-quality serum data and tandem mass spectrometry(MS/MS)data sets;relevant data collected from various types of metabolite databases.(2)After investigating and analyzing a large number of metabolomics research work,this thesis developed and integrated multiple metabolomics analysis tools,such as biochemical reaction path structure,unknown metabolite identification,ion fragmentation tree structure,etc.These tools can greatly enhance the practicality of the reference database HRSMDB.(3)Based on the collected tandem mass spectrometry data sets and related research work,this thesis proposes to use the multi-layer perceptron network to predict the molecular fingerprints of metabolites to identify unknown metabolites.According to the experimental comparison results,the molecular fingerprint prediction model designed in this thesis can effectively identify unknown metabolites.The HRSMDB provides a web page,which can be used conveniently and quickly.