SpLitteR: diploid genome assembly using TELL-Seq linked-reads and assembly graphs.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Long-read sequencing (e.g., high-fidelity or HiFi) enables contiguous genome assemblies but struggles with repeats longer than read lengths.
• Diploid genome assembly is challenged by shared long identical regions between haplomes, limiting contiguity.
• Existing methods often require additional experimental technologies to improve diploid assembly contiguity.

Purpose of the Study:

• To develop a novel tool, SpLitteR, for improving diploid genome assembly using linked-reads and assembly graphs.
• To evaluate the effectiveness of SpLitteR in phasing and scaffolding assembly graphs.
• To assess SpLitteR's performance against state-of-the-art linked-read scaffolders.

Main Methods:

• Development of the SpLitteR tool for assembly graph phasing and scaffolding.
• Utilization of barcoded linked-reads generated by TELL-Seq technology.
• Benchmarking SpLitteR against ARKS and SLR-superscaffolder using human HG002 genome and sheep gut microbiome datasets.

Main Results:

• SpLitteR achieved a 1.5-fold increase in NGA50 compared to baseline LJA assembly and other scaffolders on the human HG002 dataset.
• SpLitteR introduced no additional misassemblies in the human genome dataset.
• TELL-Seq reads were shown to facilitate phasing and scaffolding within assembly graphs.

Conclusions:

• SpLitteR effectively improves the accuracy and contiguity of diploid genome assemblies.
• The tool leverages linked-reads to bridge unresolved repeats, enhancing genome reconstruction.
• SpLitteR is available as part of the SPAdes package, offering an accessible solution for advanced genome assembly.