T2T-ONT: Diploid Telomere-to-Telomere Assembly from ONT Reads

This repository contains the analysis code and documentation for producing haplotype-resolved, near-T2T human genome assemblies using only Oxford Nanopore Technology (ONT) sequencing. Three types of ONT data are combined: ultra-long reads, HERRO-corrected reads, and Pore-C reads. The same sequencing data also yields phased CpG methylation maps and haplotype-resolved 3D chromatin contact matrices. It is the companion repository for this publication:

Single-Platform Nanopore Sequencing Enables Diploid Telomere-to-Telomere Genome Assembly and Haplotype-Resolved 3D Chromatin Maps

Caspar Gross^1,2,§, Ramya Potabattula^1,§, Fubo Cheng¹, Sarah Leuchtenberg¹, Hanna Sophie Hartung¹, Beate Kristmann¹, Elena Buena-Atienza^1,3, Nicolas Casadei^1,3, Stephan Ossowski^1,2,* and Olaf Riess^1,3,*

¹ Institute of Medical Genetics and Applied Genomics, University of Tübingen, Tübingen, Germany
² Institute for Bioinformatics and Medical Informatics (IBMI), University of Tübingen, Tübingen, Germany
³ NGS Competence Center Tübingen (NCCT), Tübingen, Germany
^§ These authors contributed equally

Repository structure

workflow/       Snakemake workflow (rules, scripts, environments)
data/           Sample and dataset configuration
assembly/       Assembly configuration and QC outputs
doc/            Analysis documentation

Documentation

• Datasets: Reference genomes, databases, and published datasets used for benchmarking
• Methods: Key analysis methods and tool configuration
• Data QC: Sequencing quality control and comparison with public datasets
• Assembly QC: QC metrics and their definitions
• Assembly Polishing: Medaka APK/ULK polishing
• TAD and Loop Analysis: 3D chromatin structure analysis
• 3D Genome: Haplotype-resolved 3D genome reconstruction
• Ancestry Analysis: Population genetics methods

Quick Start

The analysis is implemented as a Snakemake workflow. If you want to adapt or rerun the analysis, see workflow/README.md for setup and usage instructions.

Requirements

• Software: Snakemake ≥ 8.0 with Conda/Mamba
• Compute: GPU resources for HERRO error correction and Dorado basecalling, at least 500GB Ram for Assembly, minimum 64 Threads
• Sequencing data: 3× ultra-long (UL) ONT flowcells and 1× Pore-C flowcell per sample

Basic usage

# Configure samples in data/datasets.yml and assembly/assemblies.yml

snakemake --snakefile workflow/Snakefile \
    --use-conda all_samples

License

This project is licensed under the MIT License — see LICENSE for details.