Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Bioactive peptide prediction method based on deep convolutional neural network

A bioactive peptide, neural network technology, applied in neural learning methods, biological neural network models, neural architectures, etc.

Pending Publication Date: 2022-04-12
FUQING BRANCH OF FUJIAN NORMAL UNIV
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, to date, little effort has been devoted to developing methods to accurately predict AAP

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
  • Bioactive peptide prediction method based on deep convolutional neural network
  • Bioactive peptide prediction method based on deep convolutional neural network
  • Bioactive peptide prediction method based on deep convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0047] 1. Benchmark comparison between AAPred-CNN and existing prediction methods

[0048] Experimental setup. To evaluate the effectiveness of AAPred-CNN, benchmark datasets (including S main and S NT15 ) for independent testing. Note that independent testing can assess model generalization when given unseen (new) samples, which is sufficient for model comparison. We report the performance comparison and analysis of AAPred-CNN with the state-of-the-art models. image 3 The experimental results are summarized.

[0049] AAPred-CNN outperforms the current state-of-the-art models. like image 3 As shown, in the main dataset, AAPred-CNN performs 2.58%, 2.75%, 9.06%, 5.77% higher than TargetAntiAngio in terms of ACC, BACC, SP, and MCC, respectively. Similarly, in the NT15 dataset, AAPred-CNN outperforms the AntiAngioPred method by 5.00%, 5.68%, 36.84% and 13.69% on ACC, BACC, SP and MCC, respectively. Both comparative results show the comprehensive advantages of the propose...

Embodiment 2

[0051] 2. Research on the influence of AAPred-CNN hyperparameters on performance

[0052] Experimental setup. To investigate which factors play an important role in AAPred-CNN, we selected hyperparameters such as learning rate, batch size, and embedding dimension to study the impact on the performance of AAPred-CNN. For convenience, we only select commonly used candidate hyperparameters for study. Specifically, the learning rate is chosen from 0.0005 to 0.0030, and the batch size is chosen from 16 to 128. Both the number of filters and the embedding dimension range from 16 to 512. Compare the result as Figure 4-5 shown.

[0053] For the two evaluation datasets, the performance of AAPred-CNN is relatively stable when the hyperparameters are within a certain range. from Figure 4-5 It can be seen that when the learning rate is about 0.0001, the batch size is about 32, the number of filters is about 128, and the embedding dimension is about 128, the model can achieve the b...

Embodiment 3

[0055] 3. Research on the influence of training data volume on model performance

[0056] Experimental setup. As mentioned above, it is counterintuitive that deep learning models achieve superior performance despite only being given a few training samples. Therefore, it is necessary to investigate to what extent the performance of deep learning models depends on the amount of data in order to further discuss why deep learning models can work well in such few example scenarios. for S main and S NT15 . We keep the test set unchanged and use different proportions of the training set to build the evaluation model. Specifically, from 0 to 100%, every 10% is selected as an interval to randomly select the corresponding proportion of the training set. The results of the comparative experiments are as Figure 6-7 shown.

[0057] The larger the proportion of training samples, the better the model performance. Figure 6-7 The comparison results in [20] show that the overall perfo...

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 relates to a bioactive peptide prediction method based on a deep convolutional neural network, which processes an original amino acid residue sequence through one-dimensional convolution of filters with different sizes, extracts useful features, optimizes a model through optimization model output and label cross entropy loss, and is specially designed for AAP mining and prediction. The model, namely the AAPred-CNN is based on an embedding technology but not based on features of feature engineering and manual design, data sets used for training and testing of the AAPred-CNN are a disclosed data set main and an NT15 data set, the classic deep learning algorithm TextCNN is combined with the embedding technology, residue tendency analysis and the like are applied to the prediction problem of the anti-angiogenic peptide for the first time, and the prediction accuracy of the anti-angiogenic peptide is improved. A classifier AAPred-CNN with excellent performance is designed for mining and predicting the AAP, and the AAPred-CNN is based on an embedding technology, does not depend on feature engineering, and can extract useful information from a pure amino acid residue sequence in a self-adaptive mode and is used for predicting whether polypeptide has anti-angiogenesis functional activity or not.

Description

technical field [0001] The invention belongs to the field of bioinformatics technology. For the first time, the classic deep learning algorithm TextCNN is applied to the prediction of anti-angiogenic peptide (AAP, anti-angiogenic peptide) in combination with the analysis of residue tendency, and it is designed for the mining and prediction of AAP. Classifier AAPred-CNN with excellent performance. AAPred-CNN is based on embedding technology and does not rely on feature engineering. It can extract useful information from pure amino acid residue sequences in an adaptive manner, and can be used to predict whether the polypeptide has anti-angiogenic functional activity, especially involving depth-based volumes. Bioactive Peptide Prediction Method Based on Integrated Neural Networks. Background technique [0002] With the rapid growth of post-genomic data, peptide sequences are widely available in various databases, facilitating the research of therapeutic peptides. For example,...

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 Applications(China)
IPC IPC(8): G16B40/20G16B45/00G06N3/04G06N3/08
Inventor 林常航
Owner FUQING BRANCH OF FUJIAN NORMAL UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com