Unlock instant, AI-driven research and patent intelligence for your innovation.

A Protein Classification Method Based on Active Semi-Supervised Graph Neural Networks

A neural network and classification method technology, applied in neural learning methods, biological neural network models, proteomics, etc., can solve the problem of scarcity of labeled protein samples, achieve the effect of reducing classification costs and improving training effects

Active Publication Date: 2022-07-19
SHANXI UNIV +1
View PDF12 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the deficiencies in the above-mentioned prior art, and propose a protein classification method based on active semi-supervised graph neural network, which is used to solve the problem of scarcity of labeled protein samples in the existing classification methods

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Protein Classification Method Based on Active Semi-Supervised Graph Neural Networks
  • A Protein Classification Method Based on Active Semi-Supervised Graph Neural Networks
  • A Protein Classification Method Based on Active Semi-Supervised Graph Neural Networks

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0036] Attached below figure 1 The implementation steps of the present invention are further described.

[0037] Step 1, generate protein training set and test set.

[0038] Randomly select at least 1000 protein samples containing at least 50 species to form a sample set, and model each protein sample in the sample set to obtain the graph structure of the protein sample including polygonal structure and single-chain structure, and all graph structures are composed of Training set.

[0039] Each protein sample in the 1000 protein sample composition sample set includes its internal atoms, the connection relationship between atoms, the feature vector matrix of atoms and the class label of the protein.

[0040] The steps for modeling each protein sample in the sample set are as follows:

[0041] Step 1, represent each atom of each protein sample in the sample set as a node, and number each node from 0;

[0042] In step 2, the atomic connection relationship of each protein samp...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a protein classification method based on an active semi-supervised graph neural network. The steps are: (1) generating protein training sets and test sets; (2) constructing a graph neural network; (3) training a graph neural network; (4) predicting unlabeled protein samples. The invention overcomes the high cost of obtaining a large number of labeled protein samples as a training set in the prior art, and the scarcity of labeled protein samples will lead to poor model training effect. The samples are pseudo-labeled to expand the training set, so that the present invention has shorter processing time and greater space utilization when dealing with a large number of unlabeled proteins.

Description

technical field [0001] The invention belongs to the technical field of physics, and further relates to a protein classification method based on an active semi-supervised graph neural network in the technical field of image classification. The present invention can extract the attribute features of proteins from the structure and molecular node attributes of the protein graph through an active semi-supervised graph neural network, and classify proteins according to the attribute features, such as judging whether a protein is an antibody protein. Background technique [0002] Proteins, as a kind of non-Euclidean data, can be naturally represented by a graph structure, that is, by representing protein molecules as a set of objects (nodes) and their relationships (edges). In routine protein classification work, it is usually necessary to pass a series of biological experiments to determine the properties of proteins, such as determining whether a protein is an antibody protein. ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G16B20/00G06N3/04G06N3/08
CPCG16B20/00G06N3/08G06N3/045
Inventor 解宇解子璇吕圣泽鱼滨张琛
Owner SHANXI UNIV