Aktivno učenje za detekciju rijetkih epigenetskih modifikacija

undergraduate thesis

• Title (english): Active Learning for Detecting Sparse Epigenetic Modifications
• Author: Matija Pintarić
• Mentor: Mile Šikić (mentor)
• Committee member: Mile Šikić (predsjednik povjerenstva)
• Committee member: Rafael Josip Penić (član povjerenstva)
• Committee member: Filip Tomas (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Epigenetske modifikacije su vrsta modifikacija DNK koje ne mijenjaju sekvencu baza i
reverzibilne su. Pokazalo se da su one važan faktor u mnogim biološkim procesima. Ko-
rištenje modela dubokog učenja za detekciju epigenetskih modifikacija analizom podataka
dobivenih pomoću sekvenciranja temeljenog na nanoporama pokazalo se kao superiorni
pristup. Međutim, ovi modeli loše izvode detekciju rijetkih modifikacija zbog disbalansiranog skupa podataka. Glavni cilj ovog rada bio je koristiti metode aktivnog učenja stvorene za
ublažavanje disbalansa klasa i poboljšanje detekcije rijetkih modifikacija. Model koji je korišten u ovom radu je Rockfish, model dubokog učenja dizajniran za detekciju modifikacija.
Isprobane su i evaluirane tri različite metode.

Abstract (english)

Epigenetic modifications are a type of DNA modifications that do not alter the base sequence and are reversible. They have shown to be an important factor in many biological
processes. Usage of deep learning models for detection of epigenetic modification by analyzing data obtained through nanopore-based sequencing has been shown as the superior
approach. However, these models perform poorly in detecting rare modifications, because
of the disbalanced dataset. The main goal of this thesis, was to use active learning methods
created to mitigate class imbalance and improve detection of rare modifications. The model
used for this thesis is Rockfish, a deep learning model designed for modification detecting.
Three different active learning methods were tried and evaluated.

• Keywords: Bionformatika; epigenetika; Rockfish; duboko učenje; aktivno učenje
• Keywords (english): bionformatics; epigenetics; Rockfish; deep learning; active learning
• Language: english
• URN:NBN: urn:nbn:hr:168:110135
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-10-11 10:09:23