izpis_h1_title_alt

Nov način optimizacije rabe kodonov z uporabo strojnega učenja na osnovi struktur proteinov
ID Špendl, Martin (Author), ID Gunčar, Gregor (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (5,48 MB)
MD5: 25AAEB798E0F4651E362B8BC149C4537

Abstract
Izražanje sintetičnih genov v gostiteljskem organizmu zahteva prilagojeno rabo kodonov za višje izkoristke izolacije. Kljub temu da zaporedje kodonov vpliva na zvijanje proteinov, dosedanji algoritmi za optimizacijo rabe kodonov ne upoštevajo strukture proteina. S simulacijo premikanja ribosomov po mRNA lahko za dano zaporedje kodonov napovemo čas translacije na nivoju kodona. Razlaga upočasnjene translacije na podlagi strukture proteina omo\-goča napoved časa translacije za poljuben protein. Pokazali smo, da zastajanje ribosomov na mRNA povzročajo območja kodonov s povprečno višjim časom translacije. Ta so okvirno dolga 7 kodonov in povzročijo zastajanje še vsaj dveh ribosomov navzgor. Del mest zastojev ribosomov sovpada s povezovalnimi regijami, mi pa smo odkrili dodatna zastajanja, povezana z izhodom ostankov hidrofobnega jedra iz ribosoma. Iz značilk simuliranega kotranslacijskega zvijanja smo uspešno ustvarili prvi napovedni model časa translacije iz strukture proteina. Napoved veznega člena med strukturo proteina in zaporedjem kodonov omogoča optimizacijo rabe kodonov na osnovi strukture, česar dosedanje metode ne upoštevajo. Algoritem je ključnega pomena za učinkovito izražanje rekombinantnih in de novo proteinov, npr. bioloških zdravil, kjer poleg visokega izkoristka želimo pravilno zvit protein, ki je varen za uporabo.

Language:Slovenian
Keywords:raba kodonov, čas translacije, zvijanje proteinov, stojno učenje, TASEP
Work type:Master's thesis/paper
Typology:2.09 - Master's Thesis
Organization:FKKT - Faculty of Chemistry and Chemical Technology
Year:2022
PID:20.500.12556/RUL-140008 This link opens in a new window
COBISS.SI-ID:129954307 This link opens in a new window
Publication date in RUL:09.09.2022
Views:1041
Downloads:101
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:A novel approach to codon usage optimization with machine learning based on protein structure
Abstract:
Synthetic gene expression in a host organism requires adjusted codon usage for higher isolation yields. Even though codon sequence influences protein folding, current codon usage optimization algorithms do not consider protein structure. By simulating ribosome movement on mRNA for a given codon sequence, we can predict translation time with a precision of a codon. Explanation of slowed translation through protein structure enables us to predict translation time for an arbitrary protein. We showed that ribosomal stalling on mRNA is driven by codon regions with higher average translation time. These are approximately 7 codons wide and result in stalling of at least two other ribosomes upstream. A portion of ribosome stalling sites overlaps with linking regions, while we discovered more stalling sites related to hydrophobic core residues exiting the ribosome. Utilizing features from simulated co-translational folding, we successfully created the first predictive model for translation time using protein structure. Predicting the connection between protein structure and codon sequence enables codon usage optimization based on structure, which is not considered in current methods. The algorithm is of key importance for the efficient expression of recombinant and de novo proteins, i.e. biological medicinal products, where high yield must be accompanied by correct protein folding for safe usage.

Keywords:codon usage, translation time, protein folding, machine learning, TASEP

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back