Analysis of new long-read sequencing data

• Main Author: Phoa, Yohanes Alfredo
• Other Authors: Kiah Han Mao
• Format: Final Year Project
• Language: English
• Published: 2019
• Subjects: DRNTU::Science::Mathematics::Applied mathematics::Simulation and modeling; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Pattern recognition
• Online Access: http://hdl.handle.net/10356/77170
• Institution: Nanyang Technological University

The rapid development of powerful high throughput sequencing technologies has enabled us to gain valuable insights into the complexities of a human transcriptome. In recent years, Oxford Nanopore has developed a new technology that can take RNA directly as the sequencing input and generates long reads. In this thesis, we are using nanopore reading results from synthetic RNA samples and employ machine learning based approaches to identify patterns that distinguish signals from modified RNA readings from the unmodified counterpart. Firstly, we performed explorations of our dataset using a statistical test. We then proposed a simple baseline algorithm that learns the distinguishing features between unmodified strands and unmodified strands. Finally, we proposed a novel method on detecting anomalies by sequence labeling using deep learning.