Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods

碩士 === 逢甲大學 === 資訊工程學系 === 106 === An allergic reaction is an overreaction that our body's immune system misinterprets some otherwise harmless substances as a threat to our body. The substances that can cause allergic reactions are called allergens. At present, studies on allergic proteins are...

Full description

Bibliographic Details
Main Authors: SU.CHING-TING, 蘇郅挺
Other Authors: YU,CHIN-SHENG
Format: Others
Language:zh-TW
Published: 2018
Online Access:http://ndltd.ncl.edu.tw/handle/e3wxw3
id ndltd-TW-106FCU00392041
record_format oai_dc
spelling ndltd-TW-106FCU003920412019-06-27T05:28:34Z http://ndltd.ncl.edu.tw/handle/e3wxw3 Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods 經由整合機器學習方法篩選n-peptide特徵進行蛋白質序列的過敏性調查 SU.CHING-TING 蘇郅挺 碩士 逢甲大學 資訊工程學系 106 An allergic reaction is an overreaction that our body's immune system misinterprets some otherwise harmless substances as a threat to our body. The substances that can cause allergic reactions are called allergens. At present, studies on allergic proteins are almost always based on predictions. Data sets are created using known allergen proteins and non-allergenic proteins. After feature extraction, prediction models are established through machine learning methods, followed by unknown proteins. Sequences can be classified using the previously constructed predictive model. This paper builds on the future analysis of the forecast results. For further research, we use the SVM (Support Vector Machine) to integrate the first-level forecast results into the second-level forecast model. The predicted results are as follows (test set results SE = 70.9, ACC = 96.2%, SP = 99.1%, PR = 90%, MCC = 0.78). (Independent test set results SE = 73.0%, ACC = 96.4%, SP = 99.1%, PR = 90.3%, MCC = 0.79) Based on the results of this prediction, we analyzed the allergen sequence and returned the final predicted result to the original protein sequence, and we hope to obtain the analysis results related to the criticality of the allergen protein. YU,CHIN-SHENG 游景盛 2018 學位論文 ; thesis 34 zh-TW
collection NDLTD
language zh-TW
format Others
sources NDLTD
description 碩士 === 逢甲大學 === 資訊工程學系 === 106 === An allergic reaction is an overreaction that our body's immune system misinterprets some otherwise harmless substances as a threat to our body. The substances that can cause allergic reactions are called allergens. At present, studies on allergic proteins are almost always based on predictions. Data sets are created using known allergen proteins and non-allergenic proteins. After feature extraction, prediction models are established through machine learning methods, followed by unknown proteins. Sequences can be classified using the previously constructed predictive model. This paper builds on the future analysis of the forecast results. For further research, we use the SVM (Support Vector Machine) to integrate the first-level forecast results into the second-level forecast model. The predicted results are as follows (test set results SE = 70.9, ACC = 96.2%, SP = 99.1%, PR = 90%, MCC = 0.78). (Independent test set results SE = 73.0%, ACC = 96.4%, SP = 99.1%, PR = 90.3%, MCC = 0.79) Based on the results of this prediction, we analyzed the allergen sequence and returned the final predicted result to the original protein sequence, and we hope to obtain the analysis results related to the criticality of the allergen protein.
author2 YU,CHIN-SHENG
author_facet YU,CHIN-SHENG
SU.CHING-TING
蘇郅挺
author SU.CHING-TING
蘇郅挺
spellingShingle SU.CHING-TING
蘇郅挺
Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods
author_sort SU.CHING-TING
title Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods
title_short Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods
title_full Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods
title_fullStr Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods
title_full_unstemmed Allergenicity Investigation Through n-peptide Based Features on Integrated Machine Learning Methods
title_sort allergenicity investigation through n-peptide based features on integrated machine learning methods
publishDate 2018
url http://ndltd.ncl.edu.tw/handle/e3wxw3
work_keys_str_mv AT suchingting allergenicityinvestigationthroughnpeptidebasedfeaturesonintegratedmachinelearningmethods
AT sūzhìtǐng allergenicityinvestigationthroughnpeptidebasedfeaturesonintegratedmachinelearningmethods
AT suchingting jīngyóuzhěnghéjīqìxuéxífāngfǎshāixuǎnnpeptidetèzhēngjìnxíngdànbáizhìxùlièdeguòmǐnxìngdiàochá
AT sūzhìtǐng jīngyóuzhěnghéjīqìxuéxífāngfǎshāixuǎnnpeptidetèzhēngjìnxíngdànbáizhìxùlièdeguòmǐnxìngdiàochá
_version_ 1719211955399426048