Translation Initiation Site Prediction in Arabidopsis thaliana Using Synthetic Datasets and Black-Box Models

Bachelor's dissertation in Center of Biotech Data Science, Ghent University Global Campus

Yunseol Park

Counselors: Espoir Kabanga, Jasper Zuallaert

Supervisors: Arnout Van Messem, Wesley De Neve

Description

The main goals of the research effort presented in this project are as follows:

• Identify meaningful features of the TIS prediction model.
• Compare the true black-box model with the synthetic black-box model.
• Investigate the effect of noisy data on the TIS prediction.

In order to achieve these goals, the following steps are taken:

1. Generate the TIS synthetic dataset.
2. Train the model with real and synthetic data (A. thaliana).
3. Compare the results of the models trained on real and synthetic data.
4. Perform feature analysis.
5. Train the prediction model with noisy data.