Method for screening diabetes markers from body fluid metabonome profile

Abstract

The invention discloses a method for screening diabetes markers from body fluid metabonome profile. Fluid body metabolites are analyzed by using a gas phase chromatography mass spectrometer to obtain a metabonome profile, a double cross validation screening model is set up to analyze the metabonome profiles of a diabetes patient and a healthy person, and screening diabetes-specific markers. The double cross validation screening model classifies the metabonome profile as a model set and a test set in the outer layer cross validation, the model set is classified as a training set and a validation set in the inner layer cross validation, partial least squares-discriminant analysis (PLSDA) models are set up in the outer layer and the inner layer respectively, modeling process and preliminary model optimization are performed in the inner layer, the classification predication ability of the model is examined in the outer layer and ultimate model parameters are confirmed, and the markers in the metabonome profile can be obtained by using an ultimate screening model. The method can realize efficient and accurate screening for the diabetes-related markers, and has the advantages that the accuracy rate is high, the erroneous rate is small, the parameters of the screening model are optimized automatically and a data preprocessing method can be optimized. The validation of the discriminant analysis indicates that, the average accuracy of the predication reaches 98.3% and the positive detection rate reaches 100%. The method disclosed by the invention brings a prospect of high throughout screening for the diabetes markers and clinical applications.

Description

technical field [0001] The invention relates to the fields of analytical chemistry, medicine and the like, and is a new method for screening small molecular metabolic markers of diabetes by performing metabolome profile analysis on body fluids based on analytical chemistry methods and metabolomics technology. Background technique [0002] Diabetes has increasingly become one of the diseases that seriously endanger human health. It is estimated that by 2010, the number of diabetic patients in the world will reach 220 million, and by 2025 this number will increase to 300 million. Diabetes is a comprehensive metabolic disease characterized by high blood sugar caused by various complex factors. Its main pathological basis is the disorder of human metabolism such as sugar, protein, and fat caused by hypofunction of pancreatic islets or insulin resistance. At present, the detection and diagnosis of diabetes mainly focus on the detection of a few clinical indicators such as blood s...

AI Technical Summary

Problems solved by technology

However, since the modeling process also involves the interpretation of classification information, if the Q2 value is large and the model's ability to explain classification information is poor, the modeling results are not reliable at this time

[0009] 3. The preprocessing method of the original data has a great influence on the final model result, and different types of data often require different types of data preprocessing methods 1

The lack of efficient and robust integrated methods limits the application of double cross-validation methods, especially in metabolome screening

Images