In silico predikcija aktivnosti peptida

undergraduate thesis

• Title (english): In silico prediction of peptide bioactivity
• Author: Mihovil Pupačić
• Mentor: Antonio Starčević (mentor)
• Granter: University of Zagreb; Faculty of Food Technology and Biotechnology; (Department of Biochemical Engineering); (Section for Bioinformatics); Zagreb
• Defense date and country: 2024-09-16, Croatia
• Scientific / art field, discipline and subdiscipline: BIOTECHNICAL SCIENCES; Biotechnology

Abstract

Bioinformatički alati imaju sve širu primjenu u području analize proteina, njihovih sekvenci te aktivnosti. Na temelju sastava polipeptidnog lanca, međusobnih interakcija aminokiselina u njemu te ostalih parametara, razni bioinformatički alati utemeljeni na elementima strojnog učenja omogućuju pobližu predikciju bioaktivnosti i toksičnosti proteina. U ovom radu nasumično je odabrano 100 proteina iz ToxProt baze podataka te je svakom pronađen homolog BLAST pretraživanjem. Izabrani homolozi su proteini koji još nisu klasificirani unutar ToxProt baze podataka te pokazuju što nižu E-vrijednost jer to ukazuje na visoki stupanj homologije. Za svaki od ukupno 200 proteina je provedena predikcija bioaktivnosti pomoću bioinformatičkog alata PeptideRanker te predikcija toksičnosti pomoću ToxinPred2. Proteini sa visokim stupnjem homologije i sličnim rezultatima predikcija su dodatno 3D generirani AlphaFold alatom, čime je njihova sličnost potvrđena. S obzirom na homolognost sekvence te 3D strukture proteina i njegovog homologa bilo je moguće predvidjeti aktivnost homologa kojemu je do sada funkcija još bila nepoznata.

Abstract (english)

Bioinformatics tools are increasingly applied in the analysis of proteins, their sequences, and activities. Based on the composition of polypeptide chains, amino acid interactions, and other parameters, machine learning-based bioinformatics tools enable predictions of protein bioactivity and toxicity. In this study, 100 proteins were randomly selected from the ToxProt database, and homologs were identified using BLAST search. These homologs, not yet classified in ToxProt, exhibited low E-values, indicating a high degree of homology. Bioactivity was predicted for all 200 proteins using PeptideRanker, and toxicity using ToxinPred2. Proteins with high homology and similar prediction results were further 3Dmodeled with AlphaFold, confirming their similarity. Based on sequence homology and 3D structure comparison, it was possible to predict the activity of homologs whose functions were previously unknown.

• Keywords: bioaktivnost; toksin; homologija; ToxProt; BLAST; PeptideRanker; ToxinPred2; AlphaFold
• Keywords (english): bioactivity; toxin; homology; ToxProt; BLAST; PeptideRanker; ToxinPred2; and AlphaFold
• Language: croatian
• URN:NBN: urn:nbn:hr:159:478383
• Study programme: Title: Biotechnology; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka biotehnologije (sveučilišni/a prvostupnik/prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka biotehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-11-07 14:08:27