Triple quadrupole mass spectrometer spectrogram classification method based on Hopfield model

Abstract

The invention provides a triple quadrupole mass spectrometer spectrogram classification method based on a Hopfield model, and the method comprises the following steps: S1, improving the Hopfield model based on the feature information of various target peaks; s2, extracting points within a certain distance range on the left side and the right side of the highest point of the target peak, and subjecting the points to normalization processing so as to serve as initial state vectors of the improved Hopfield model; and S3, generating attractors and attraction basins of the Hopfield model based on the priori training set, and when the data of the peak to be identified falls into the attraction basins of the corresponding attractors, finally converging to the attractors along with recursive calculation, the attractors being a so-called type identifier. The Hopfield model is applied to spectrogram classification application of the triple quadrupole mass spectrometer, the method is a nonlinear recurrent neural network, and compared with methods such as a support vector machine, the calculation complexity is low, the characteristics of a target peak do not need to be specially concerned, and the method is easy to implement. Meanwhile, the number of neurons of the Hopfield model can be flexibly adjusted, so that scanning scenes with different synchronization lengths can be dealt with.

Description

technical field [0001] The invention belongs to the technical field of mass spectrometers, in particular to a Hopfield model-based spectrum classification method for triple quadrupole mass spectrometers. Background technique [0002] Mass spectrometry is an important tool for qualitative and quantitative analysis of compounds at present. There are many types and are widely used in school laboratories, hospitals, research institutes and pharmaceutical companies. The triple quadrupole mass spectrometer is also one of them. The triple quadrupole mass spectrometer provides a cascade function to enhance the qualitative ability of the mass spectrometer. In addition, it has good quantitative ability and high MRM signal-to-noise ratio. In addition to general product ion scanning functions, it also supports product ion scanning, precursor ion scanning, and neutral loss (Neutral loss) and other modes. The triple quadrupole mass spectrometer has powerful functions, but it has many par...

AI Technical Summary

Problems solved by technology

The triple quadrupole mass spectrometer has powerful functions, but there are many parameters and tuning is difficult. Especially in automatic tuning, it is necessary to identify and distinguish target peaks and miscellaneous peaks, which has become one of the difficulties in automatic tuning.

During manual tuning, it often requires extensive experimental experience to identify the target peak, which is prone to errors

If you make a mistake, it will often lead to abnormal experimental results, and what's more, it will cause certain losses.

Images