prikaz prve stranice dokumenta Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža
Access restricted to students and staff of home institution
master's thesis
Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža
Rijeka: University of Rijeka, Faculty of Engineering, 2021. urn:nbn:hr:190:168278
Vidaković-Lipovac, Marin
University of Rijeka
Faculty of Engineering
Department of Computer Engineering
Section of Software Engineering

Institutional repository: Repository of the University of Rijeka, Faculty of Engineering

Cite this document

Vidaković Lipovac, M. (2021). Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža (Master's thesis). Rijeka: University of Rijeka, Faculty of Engineering. Retrieved from https://urn.nsk.hr/urn:nbn:hr:190:168278

Vidaković Lipovac, Marin. "Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža." Master's thesis, University of Rijeka, Faculty of Engineering, 2021. https://urn.nsk.hr/urn:nbn:hr:190:168278

Vidaković Lipovac, Marin. "Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža." Master's thesis, University of Rijeka, Faculty of Engineering, 2021. https://urn.nsk.hr/urn:nbn:hr:190:168278

Vidaković Lipovac, M. (2021). 'Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža', Master's thesis, University of Rijeka, Faculty of Engineering, accessed 25 December 2024, https://urn.nsk.hr/urn:nbn:hr:190:168278

Vidaković Lipovac M. Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža [Master's thesis]. Rijeka: University of Rijeka, Faculty of Engineering; 2021 [cited 2024 December 25] Available at: https://urn.nsk.hr/urn:nbn:hr:190:168278

M. Vidaković Lipovac, "Konstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža", Master's thesis, University of Rijeka, Faculty of Engineering, Rijeka, 2021. Available at: https://urn.nsk.hr/urn:nbn:hr:190:168278

Metadata
TitleKonstruktivno pretraživanje peptida zasnovano na principu suparničkih mreža
Title (english)Constructive peptide design based on the adversarial networks principle
AuthorMarin Vidaković-Lipovac
MentorGoran Mauša (mentor)
Committee memberSandi Ljubić (predsjednik povjerenstva)
Committee memberIvan Štajduhar (član povjerenstva)
Committee memberGoran Mauša (član povjerenstva)
GranterUniversity of Rijeka
Faculty of Engineering
(Department of Computer Engineering)
(Section of Software Engineering)
Rijeka
Defense date and country2021-09-20, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
U ovom radu korištene su generativne suparničke mreže kako bi se generirale nove sekvence peptida s antimikrobijskim svojstvima. Predstavljena je priprema podataka u kojoj su se postojeći antimikrobni peptidi morali pretvoriti u one-hot vektorski zapis. Klasičan model generatora izmijenjen je tako što mu je zadnji sloj bio podijeljen u 50 skupina po 21 neuron, te je za svaku skupinu korištena omekšana softmax aktivacijska funkcija. Za model diskriminatora je korišten obični klasifikacijski model. Model je uspješno istreniran, vidljiv je porast udjela generiranih antimikrobnih peptida kroz epohe, 56% nakon 5000 epoha, te čak 86% nakon 10 000 epoha. Analiza sekvenci pokazala je da su generirani peptidi kraći od peptida korištenih za trening, te je također vidljivo da se aminokiseline alanin (A), glicin (G), lizin (K), leucin (L) pojavljuje češće, iznad 10%. Dok se aminokiseline glutaminska kiselina (E), histidin (H), metionin (M), triptofan (W), tirozin (Y) pojavljuju rjeđe, ispod 1%.
Abstract (english)
In this thesis, generative adversarial networks were used to generate new peptide sequences with antimicrobial activity. Data preparation was presented in which existing antimicrobial peptides had to be converted to a one-hot vector record. The classic model of the generator was modified, the last network layer was divided into 50 groups of 21 neurons, and a softened softmax activation function was used for each group. An ordinary classification model was used for the discriminator model. The model has been successfully trained, the increase in the share of generated antimicrobial peptides through epochs is visible, 56% after 5000 epochs, and even 86% after 10 000 epochs. Sequence analysis showed that the generated peptides were shorter than the peptides used for training, and it was also seen that the amino acids alanine (A), glycine (G), lysine (K), leucine (L) occur more frequently, above 10 %. While the amino acids glutamic acid (E), histidine (H), methionine (M), tryptophan (W), tyrosine (Y) occur less frequently, below 1%.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:190:168278
Study programmeTitle: Computer Engineering
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to students and staff of home institution
Terms of useLicence In copyright icon
Created on2021-10-15 12:33:58