prikaz prve stranice dokumenta INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION
  Download
PDF 17.14 MB
doctoral thesis
INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION
Zagreb: University of Zagreb, Faculty of Science, 2024. urn:nbn:hr:217:757529
Štancl, Paula
University of Zagreb
Faculty of Science
Department of Biology

Institutional repository: Repository of the Faculty of Science

Cite this document

Štancl, P. (2024). INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION (Doctoral thesis). Zagreb: University of Zagreb, Faculty of Science. Retrieved from https://urn.nsk.hr/urn:nbn:hr:217:757529

Štancl, Paula. "INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION." Doctoral thesis, University of Zagreb, Faculty of Science, 2024. https://urn.nsk.hr/urn:nbn:hr:217:757529

Štancl, Paula. "INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION." Doctoral thesis, University of Zagreb, Faculty of Science, 2024. https://urn.nsk.hr/urn:nbn:hr:217:757529

Štancl, P. (2024). 'INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION', Doctoral thesis, University of Zagreb, Faculty of Science, accessed 18 December 2024, https://urn.nsk.hr/urn:nbn:hr:217:757529

Štancl P. INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION [Doctoral thesis]. Zagreb: University of Zagreb, Faculty of Science; 2024 [cited 2024 December 18] Available at: https://urn.nsk.hr/urn:nbn:hr:217:757529

P. Štancl, "INFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION", Doctoral thesis, University of Zagreb, Faculty of Science, Zagreb, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:217:757529

Metadata
TitleINFLUENCE OF GENOMIC CHARACTERISTICS ON THE OPTIMIZATION OF MODELS FOR TUMOR CELL-OF-ORIGIN PREDICTION
Title (croatian)UTJECAJ GENOMSKIH KARAKTERISTIKA NA OPTIMIZACIJU MODELA ZA PREDVIĐANJE ISHODIŠNE STANICE TUMORA
AuthorPaula Štancl
MentorRosa Karlić (mentor)
Committee memberKristian Vlahoviček (predsjednik povjerenstva)
Committee memberInga Urlić (član povjerenstva)
Committee memberPetar Ozretić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Science
(Department of Biology)
Zagreb
Defense date and country2024, Croatia
Scientific / art field,
discipline and subdiscipline		NATURAL SCIENCES
Biology
Universal decimal classification
(UDC)		577 - Biochemistry. Molecular biology. Biophysics
Abstract
Identifying the cell-of-origin (COO) for tumors of unknown primary site is crucial for selecting effective therapies, but the heterogeneity and unique genomic profiles of these cancers make accurate prediction by machine learning models challenging. In this study, I analyzed how various genomic features influence COO prediction by using different machine learning models based on the mutational landscapes and epigenetic characteristics of normal tissues in distinct genomic features. The analysis of breast, liver, and skin melanoma cancers revealed significant differences of their mutational landscapes, and their influence on COO model accuracy. APOBEC mutations in early-replicating regions associated with kataegis were linked to poorer COO predictions in breast cancer. Including protein-coding genes and topologically associated domains can also be used for reliable COO predictions and identification of under-predicted genes with potential roles in tumorigenesis. While whole-exome sequencing and RNA-seq mutations aid COO identification, they are less accurate than whole-genome sequencing. Overall, the tumor's complex genomic landscape significantly affects COO prediction accuracy.
Abstract (english)
Određivanje ishodišne stanice tumora (engl. cell-of-origin, COO) ključno je za odabir učinkovitih terapija. Međutim, heterogenost i jedinstveni genomski profili ovih tumora čine njihovu preciznu identifikaciju pomoću modela strojnog učenja izazovnom. U ovom radu analizirala sam kako različite genomske značajke utječu na predviđanje COO koristeći različite modele strojnog učenja temeljene na mutacijskim krajolicima i epigenetskim karakteristikama normalnih tkiva. Analizom tumora dojke, jetre i melanoma kože otkrivene su značajne razlike u njihovim mutacijskim krajolicima te njihov utjecaj na točnost modela za predviđanje COO. APOBEC mutacije koje se pojavljuju u rano replicirajućim regijama i povezane sa žarištima mutacija, pokazale su lošiju točnost predviđanja COO kod tumora dojke. Geni za kodiranje proteina i topološki povezane domene također se mogu koristiti za pouzdana predviđanja COO, kao i za identifikaciju gena s premalim brojem predviđenih mutacija, koji imaju potencijalnu ulogu u tumorigenezi. Iako mutacije dobivene sekvenciranjem cijelog egzoma i transkriptoma pomažu u identifikaciji COO, one su manje točne od mutacija dobivenih sekvenciranjem cijelog genoma. Sveukupni rezultati ukazuju na to da kompleksni genomski krajolik tumora značajno utječe na točnost predviđanja COO.
Keywords
Keywords (english)
Languageenglish
URN:NBNurn:nbn:hr:217:757529
Promotion2024
ProjectNumber: IP-2019-04-9308
Title: Predviđanje ishodišnih stanica i istraživanje mehanizama razvoja raka bazirano na statističkom modeliranju
Title: A statistical modelling approach to predict the cell-of-origin and investigate mechanisms of cancer development
Acronym: PREDI-COO
Leader: Rosa Karlić
Jurisdiction: Croatia
Funder: Hrvatska zaklada za znanost
Funding stream: Research Projects
Study programmeTitle: Doctoral study
Study programme type: university
Study level: postgraduate
Academic / professional title: doktor/doktorica znanosti u području prirodnih znanosti (doktor/doktorica znanosti u području prirodnih znanosti)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2024-12-02 11:13:59