(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Prediction of antimicrobial peptides based on sequence alignment and support vector machine - pairwise algorithm /Ng Xin Yi

Prediction of antimicrobial peptides based on sequence alignment and support vector machine - pairwise algorithm_Ng Xin Yi_E3_2013_NI
Projek ini adalah tertumpu untuk mewujudkan satu kaedah baru untuk meramal peptida antimikrobial (AMP) yang mempunyai peranan penting dalam sistem imun. Kebelakangan ini, penyelidik mendapati bahawa sebilangan besar bakteria telah menjadi kebal kepada antibiotik yang sedia ada. Hal ini telah mendorong minat penyelidik untuk menjadikan AMPs sebagai antibiotik alternatif yang baru. Walaubagaimanapun, eksperimen untuk mengidentifikasi AMPs daripada jujukan protein adalah mahal dan memerlukan masa yang panjang. Jadi, alat ramalan berdasarkan kaedah pengiraan perlu direka untuk menyelesaikan masalah ini. Dalam projek ini, satu alat ramalan untuk AMP telah dibina dengan mengintegrasikan penjajaran urutan dan mesin sokongan vektor (SVM) dengan algorithma berpasangan. Dalam kaedah penjajaran urutan, pengiraan skor HSPs dijalankan untuk mendapat skor persamaan antara urutan ujian dan urutan latihan. Kelas urutan ujian ini adalah sama dengan kelas urutan latihan yang mempunyai skor HSPs yang maksimum. Oleh sebab penjajaran urutan tidak dapat menguruskan semua urutan peptida, algoritma SVM-berpasangan telah digunakan. Algoritma ini menggabungkan konsep skor persamaan berpasangan dengan SVM. Urutan telah diwakili dalam skor vektor yang dihasilkan berdasarkan konsep LZ kerumitan. SVM telah digunakan sebagai pengelas untuk mengklasifikasi urutan ini berdasarkan ciri-ciri vektor yang sedia ada. Prestasi alat ramalan ini telah dibandingkan dengan kaedah-kaedah yang sedia ada. Kaedah yang dicadangkan dalam projek ini telah menunjukkan prestasi yang terbaik berbanding dengan kaedah-kaedah lain berdasarkan aspek sensitiviti. Kaedah ini berupaya mencapai sensitiviti 88.74% dalam ujian Jackknife dan sensitiviti 87.59% dalam ujian Independent. Keputusan simulasi ini telah membukitkan bahawa kaedah ini sesuai untuk digunakan sebagai alat ramalan AMPs. ___________________________________________________________________________________ This project is concentrated on the attempt to establish a new method for predicting Antimicrobial Peptides (AMPs) that are important in the immune system. Recently, researchers have been interested in designing alternative drugs based on the AMPs because they found that a large number of bacterial strains have become resistant to the available antibiotics. However, the researchers have met obstacle on the AMPs identifying process because of the experiment to extract AMPs from protein sequences are costly and require longer time to setup. Therefore, a computational tool for AMPs prediction is needed to solve the problem. In this project, an integrated algorithm has been introduced to predict AMPs by combining sequence alignment and support vector machine-pairwise (SVM) algorithm. The sequence alignment method is performed by calculating the HSPs scores that represent the local similarity scores between test and training sequences. The classification of the test sequence depends on the class of the training sequence that has the maximum HSPs score. Due to sequence alignment cannot deal with all peptide sequences, SVM-pairwise also is applied to predict the remaining unpredictable sequences. SVM-pairwise algorithm combines the concept of pairwise similarity score and SVM. The remaining sequences are represented by feature vectors that contain pairwise similarity scores of the test sequence and training sequences. The feature vectors are used for classification using SVM as a classifier. In the comparison with other previously proposed methods, the proposed algorithm in this project shows the best result based on sensitivity measure. The sensitivity of the proposed algorithm is 88.74% for Jackknife test and 87.59% for Independent test. This indicates that the proposed algorithm in this project is suitable to be used as AMPs predictor.
Contributor(s):
Ng Xin Yi - Author
Primary Item Type:
Final Year Project
Language:
English
Subject Keywords:
antibiotics ; Antimicrobial Peptides ; alternative drugs
First presented to the public:
6/1/2013
Original Publication Date:
2/19/2020
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 72
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2020-02-19 12:04:29.6
Submitter:
Nor Hayati Ismail

All Versions

Thumbnail Name Version Created Date
Prediction of antimicrobial peptides based on sequence alignment and support vector machine - pairwise algorithm /Ng Xin Yi1 2020-02-19 12:04:29.6