prikaz prve stranice dokumenta Identification of gene signature of human inflammatory breast cancer using machine learning methods
Embargo period 2025-09-27
master's thesis
Identification of gene signature of human inflammatory breast cancer using machine learning methods
Zagreb: University of Zagreb, Faculty of Science, 2023. urn:nbn:hr:217:263476
Lovrečić, Karlo
University of Zagreb
Faculty of Science
Department of Biology

Institutional repository: Repository of the Faculty of Science

Cite this document

Lovrečić, K. (2023). Identification of gene signature of human inflammatory breast cancer using machine learning methods (Master's thesis). Zagreb: University of Zagreb, Faculty of Science. Retrieved from https://urn.nsk.hr/urn:nbn:hr:217:263476

Lovrečić, Karlo. "Identification of gene signature of human inflammatory breast cancer using machine learning methods." Master's thesis, University of Zagreb, Faculty of Science, 2023. https://urn.nsk.hr/urn:nbn:hr:217:263476

Lovrečić, Karlo. "Identification of gene signature of human inflammatory breast cancer using machine learning methods." Master's thesis, University of Zagreb, Faculty of Science, 2023. https://urn.nsk.hr/urn:nbn:hr:217:263476

Lovrečić, K. (2023). 'Identification of gene signature of human inflammatory breast cancer using machine learning methods', Master's thesis, University of Zagreb, Faculty of Science, accessed 03 May 2024, https://urn.nsk.hr/urn:nbn:hr:217:263476

Lovrečić K. Identification of gene signature of human inflammatory breast cancer using machine learning methods [Master's thesis]. Zagreb: University of Zagreb, Faculty of Science; 2023 [cited 2024 May 03] Available at: https://urn.nsk.hr/urn:nbn:hr:217:263476

K. Lovrečić, "Identification of gene signature of human inflammatory breast cancer using machine learning methods", Master's thesis, University of Zagreb, Faculty of Science, Zagreb, 2023. Available at: https://urn.nsk.hr/urn:nbn:hr:217:263476

Metadata
TitleIdentification of gene signature of human inflammatory breast cancer using machine learning methods
Title (english)Identifikacija genskog potpisa upalnog raka dojke čovjeka metodama strojnog učenja
AuthorKarlo Lovrečić
MentorRosa Karlić (mentor)
Committee memberRosa Karlić (predsjednik povjerenstva)
Committee memberZorana Grubić (član povjerenstva)
Committee memberIvana Šola (član povjerenstva)
Committee memberMaja Matulić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Science
(Department of Biology)
Zagreb
Defense date and country2023-09-27, Croatia
Scientific / art field,
discipline and subdiscipline		NATURAL SCIENCES
Biology
Abstract
Inflammatory breast cancer (IBC) is a rare subtype of breast cancer characterized by its fast progression, aggressiveness, and symptoms similar to acute breast inflammation. The key role in IBC plays its complex tumour microenvironment. Despite the extensive research, an IBC-specific gene signature remains to be discovered. A major obstacle in IBC research is the fact that most of the available genomic data is obtained using microarray technology which has its limitations. A computational deconvolution approach provides a new perspective in investigating IBC. Deconvolution methods can overcome challenges regarding tumour samples obtained using distinct microarray platforms, and the difference between ratios of tumour and stromal cells within samples from distinct datasets. In this graduation thesis, the deconvolution method CIBERSORTx was performed to obtain relative cell fractions of different cell types in IBC and non-IBC samples, and to obtain cell-type specific gene expression profiles. Statistical analysis of relative cell fractions revealed statistically significant differences in relative cell fractions of different cell types between IBC and non-IBC samples. Random forest and SVM models were trained using cell-type specific gene expression profiles and their performance was evaluated using an independent test dataset comprised of 110 samples (39 IBC, 71 non-IBC).
Abstract (croatian)
Upalni rak dojke (engl. inflammatory breast cancer, IBC) je rijedak podtip raka dojke karakteriziran brzom progresijom i simptoma nalik akutnoj upali dojke. Ključnu ulogu kod upalnog raka dojke ima kompleksan tumorski mikrookoliš. Unatoč intenzivnim istraživanjima, genski potpis specifičan za upalni rak dojke i dalje nije otkriven. Glavna prepreka u istraživanjima upalnog raka dojke su ograničenja DNA-mikročipova, kojima je dobivena većina dostupnih genomskih podataka o ovom tipu raka. Računalna dekonvolucija predstavlja novi pristup u istraživanju upalnog raka dojke. Korištenjem računalne dekonvolucije mogu se prevladati izazovi vezani uz analizu uzoraka tumora dobivenih korištenjem različitih DNA-mikročipova te usporediti razlika između omjera tumorskih i stromalnih stanica u uzorcima porijeklom iz različitih setova podataka. U ovom diplomskom radu, provedena je metoda CIBERSORTx kojom su dobiveni relativni stanični udjeli različitih staničnih tipova za uzorke tkiva upalnog i neupalnog raka dojke. Također, dobiveni su i ekspresijski profili specifični za pojedini tip stanica. Statističkom analizom utvrđeno je da postoje statistički značajne razlike u relativnim staničnim udjelima analiziranih staničnih tipova između uzoraka tkiva upalnog i neupalnog raka dojke. Modeli strojnog učenja „Random forest“ i „SVM“ istrenirani su korištenjem ekspresijskih profila dobivenih metodom CIBERSORTx. Preciznost klasifikacije istreniranih modela strojnog učenja utvrđena je na testnom setu podataka koji sadrži 110 uzoraka (39 IBC, 71 non-IBC).
Keywords
Keywords (croatian)
Languageenglish
URN:NBNurn:nbn:hr:217:263476
Study programmeTitle: Molecular Biology
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra molekularne biologije (magistar/magistra molekularne biologije)
Type of resourceText
File originBorn digital
Access conditionsEmbargoed access
Embargo expiration date: 2025-09-27
Terms of useLicence In copyright icon
Created on2023-10-17 11:30:13