The development of a high-performance nanopore methylation detection method with consideration of structural variation

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K12104
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 62010:Life, health and medical informatics-related
• Research Institution: The University of Tokyo
• Principal Investigator: ZHANG Yaozhong 東京大学, 医科学研究所, 准教授 (60817138)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: methylation / nanopore / deep learning

Outline of Final Research Achievements

In this project, we developed both model-level and pipeline-level high-performance methylation callers for nanopore sequencing data. We developed methBERT using the encoder architecture of the transformer model. In addition to signal analysis, we investigated the learning of nucleotide representation in the BERT model through pre-training. We analyzed representations for signals and nucleotides and developed a novel methylation caller based on the alignment of reads at target positions. At the pipeline level, we built a haplotype-aware and structural-variant-informed methylation detection pipeline, which we tested on both normal and tumor cells. Besides developing high-performance methylation callers, we extended our findings to whole-genome-level nucleotide sequence representation and single-cell representations using contrastive learning with biological constraints.

Free Research Field

bioinformatics

Academic Significance and Societal Importance of the Research Achievements

ゲノムシーケンシングのコストが安くなるにつれて、その利用も広がってきた。ゲノムシーケンシングデータをより迅速かつ高精度に解析することは、ヘルスケアや疾患診断において重要である。本研究では、ナノポアシーケンシングから高精度なメチル化プロファイリング解析技術を開発した。この技術により、メチル化を高速かつ高精度な検出することが可能になり、老化や疾患におけるエピジェネティックな変化を理解するために役割を果たすことが期待される。