Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera

Novosel, Bruno

prikaz prve stranice dokumenta Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera

Download
PDF 2.26 MB

master's thesis

Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera

2023. urn:nbn:hr:190:513051

Novosel, Bruno

University of Rijeka
Faculty of Engineering

Cite this document

APA 6th Edition

Novosel, B. (2023). Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera (Master's thesis). Rijeka: University of Rijeka, Faculty of Engineering. Retrieved from https://urn.nsk.hr/urn:nbn:hr:190:513051

MLA 8th Edition

Novosel, Bruno. "Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera." Master's thesis, University of Rijeka, Faculty of Engineering, 2023. https://urn.nsk.hr/urn:nbn:hr:190:513051

Chicago 17th Edition

Harvard

Novosel, B. (2023). 'Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera', Master's thesis, University of Rijeka, Faculty of Engineering, accessed 09 November 2024, https://urn.nsk.hr/urn:nbn:hr:190:513051

Vancouver

Novosel B. Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera [Master's thesis]. Rijeka: University of Rijeka, Faculty of Engineering; 2023 [cited 2024 November 09] Available at: https://urn.nsk.hr/urn:nbn:hr:190:513051

IEEE

B. Novosel, "Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera", Master's thesis, University of Rijeka, Faculty of Engineering, Rijeka, 2023. Available at: https://urn.nsk.hr/urn:nbn:hr:190:513051

Cite this item: https://urn.nsk.hr/urn:nbn:hr:190:513051

Please login to the repository to save this object to your list.

Metadata

Title	Analiza sličnosti peptida temeljena na latentnom prostoru autoenkodera
Title (english)	Peptide similarity analysis based on autoencoder latent space
Author	Bruno Novosel
Mentor	Goran Mauša (mentor)
Committee member	Mladen Tomić (predsjednik povjerenstva)
Committee member	Sandi Ljubić (član povjerenstva)
Committee member	Goran Mauša (član povjerenstva)
Granter	University of Rijeka Faculty of Engineering Rijeka
Defense date and country	2023, Croatia
Scientific / art field, discipline and subdiscipline	TECHNICAL SCIENCES Computing
Abstract	Unatoč brzom napretku medicine i znanosti, mnoge bolesti i dalje predstavljaju veliku prijetnju ljudskom zdravlju. U tom kontekstu, peptidi su se istaknuli kao obećavajuće novo rješenje pri liječenju. Razne metode strojnog učenja pokazale su se vrlo učinkovitima za otkrivanje novih peptida sa željenim svojstvima. U ovom radu proveli smo istraživanje o primjeni varijacijskih autoenkodera za analizu sličnosti peptida s antimikrobnim i antiviralnim svojstvima kako bismo pokušali bolje shvatiti vezu između njihovih sekvenci i fizikalno-kemijskih značajki. Kao krajnji rezultat dobiven je VAE koji uspješno uči smislene prikaze peptida u dvodimenzionalnom latentnom prostoru, koji je uspješan u rekonstrukciji peptida iz latentnog prostora i koji se pokazao kao moćan pristup istraživanju strukturnih veza među peptidima. Potreban je dodatan trud i ulaganja u slična istraživanja jer imaju velik potencijal za ostvarenje napretka u medicini i poboljšanje kvalitete ljudskog života.
Abstract (english)	Despite the quick progress in medicine and science, there are still many diseases that pose a threat to human health. In this context, peptides stood out as a new promising treatment solution. Various machine learning methods have proven very effective in discovering new peptides with desired properties. In this paper, we conducted research on the application of variational autoencoders for similarity analysis of peptides with antimicrobial and antiviral properties so we could better understand the connection between their sequences and physicochemical properties. As the end result we got a variational autoencoder that successfully learns meaningful representations of peptides in a two-dimensional latent space, that is successful in reconstructing peptides from latent space and has proven to be a powerful approach to researching structural bonds between peptides. Additional effort and investment in similar research is needed because of its great potential for achieving progress in medicine and improving the quality of human life.
Keywords
Keywords (english)
Language	croatian
URN:NBN	urn:nbn:hr:190:513051
Study programme	Title: Computer Engineering Study programme type: university Study level: graduate Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resource	Text
File origin	Born digital
Access conditions	Open access
Terms of use
Created on	2023-09-12 14:08:50