Predikcija topivosti u vodi kemijskih spojeva korištenjem kvantitativnih modela strukture i aktivnosti

undergraduate thesis

• Title (english): Prediction of water solubility of chemical compounds using quantitative structure-activity models
• Author: Katarina Pešić
• Mentor: Mihaela Vranić (mentor)
• Committee member: Mihaela Vranić (predsjednik povjerenstva)
• Committee member: Damir Pintar (član povjerenstva)
• Committee member: Luka Humski (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-07-02, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Uloga kemoinformatike u znanosti i industriji ne može se zanemariti. Kvalitetna analiza
ponašanja molekula ključan je korak u, primjerice, farmaciji. Ocjena dobrote molekule može
se donijeti na osnovu i njezine topivosti u vodi, jer ona direktno utječe na njezino ponašanje
u tijelu i u okolišu. Pokazuje se da se topivost vode može dobro modelirati linearnim i
nelinearnim kombinacijama karakteristika molekula, zvanih deskriptori, a posebno na novo
i modelima strojnog učenja. U ovom radu korištena su tri modela na tri različita skupa
podataka i pokazuje se da je moguće, kroz ispravnu provedbu, nekim modelima strojnog
učenja kvalitetno predvidjeti topivosti molekula što je bitan pokazatelj toga da strojno učenje
ima veliku ulogu i može ostati relevantno u području kemoinformatike.

Abstract (english)

The role of chemoinformatics in various fields of industry and science cannot be overlooked.
Analysis of molecular behavior is a key step in, e.g., pharmacy. Goodness of fit of one
particular molecule can be described through its water solubility, as it directly affects its
behavior in the human body, as well as in the environment. It can be shown that water
solubility may be well modeled through linear and non-linear combinations of other
molecule properties, also called molecular descriptors, especially when paired with machine
learning. In this thesis three distinct machine learning models were trained on three distinct
data sets and it was shown that, when operated in the right way, it is possible to achieve
trustworthy predictions of water solubility with machine learning. That serves as a
confirmation of the role machine learning is taking in chemoinformatics, and why it is on a
path to stay relevant in the field for the time to come.

• Keywords: Deskriptori molekula; QSAR; Topivost; Modeli strojnog učenja; Kemoinformatika
• Keywords (english): Molecular descriptors; QSAR; Solubility; Machine learning models; Chemoinformatics
• Language: croatian
• URN:NBN: urn:nbn:hr:168:356717
• 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: 2024-10-12 21:50:04