DegrDL: Pendekatan Deep Learning Untuk Memprediksi Degradasi Rna Pada Vaksin mRNA Covid-19

Jerry, Lasama (2022) DegrDL: Pendekatan Deep Learning Untuk Memprediksi Degradasi Rna Pada Vaksin mRNA Covid-19. Undergraduate Thesis thesis, Institut Teknologi Telkom Purwokerto.

	Text Cover.pdf Download (1MB)
	Text Abstract.pdf Download (43kB)
	Text BAB I.pdf Download (47kB)
	Text BAB II.pdf Download (125kB)
	Text BAB III.pdf Download (448kB)
	Text BAB IV.pdf Restricted to Registered users only Download (260kB) | Request a copy
	Text BAB V.pdf Download (41kB)
	Text Daftar Pustaka.pdf Download (69kB)

Abstract

As the preliminary report, the messenger RNA (mRNA) vaccine for COVID-19 showed all participants’ immune responses without trial-limiting safety concerns with high efficacy. However, there is a major setback in the RNA vaccine solu tion, which is the degradation of RNA that results in instability and renders mRNA useless when certain modifications were applied. Contrasted to conventional vac cines such as flu shots that could be packaged in disposable syringes and shipped under refrigeration around the world, currently, mRNA vaccines cannot be treated the same way unless there is a breakthrough in the development of the super stable RNA. To understand RNA molecules better, DasLab from Stanford Biochemistry Department developed a scientific discovery game platform that challenges play ers to solve scientific problems such as RNA design called Eterna. Data gathered from experimentations in Eterna that contains RNA Sequence, Structure, Predicted Loop Type, and also the base-pair-probabilities later was used in this research to find out the stability of the RNA molecules and which parts of the molecule are prone to degradation by leveraging Deep Learning techniques especially 1-D Con volutional Neural Network (CNN) and Bidirectional Gated Recurrent Unit (GRU) and Long-short Term Memory (LSTM). Testing results from several model com binations shows 1D-CNN Model achieved 0.184 MCRMSE and 0.178 Validation MCRMSE and converges in 34 minutes followed by the LSTM-LSTM model that achieved 0.185 MCRMSE and 0.191 Validation MCRMSE. Based on the results, it could be inferred as the 1D-CNN model is better performing for predicting RNA degradation and overall stability. Keywords: mRNA Vaccine, COVID-19, Deep Learning

Actions (login required)

View Item