Integration of string and de Bruijn graphs for genome assembly

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• String and de Bruijn graphs are standard in genome assembly.
• No single assembler excels across all datasets.
• String graphs resolve repeats; de Bruijn graphs handle error-prone regions.

Purpose of the Study:

• To develop a novel genome assembler, StriDe.
• To leverage advantages of both string and de Bruijn graph models.
• To improve genome assembly contiguity and accuracy.

Main Methods:

• Developed StriDe, a hybrid assembler.
• Adaptive read decomposition in error-prone regions.
• Paired-end read extension using FM-index for long reads.
• Constructed assembly graph from processed reads.
• Designed and improved core assembler components.
• Fully parallelized the assembler.

Main Results:

• StriDe shows contiguity comparable to leading assemblers.
• Achieved high assembly accuracy on short-read and long-read datasets.
• Demonstrated robust performance across benchmark datasets.

Conclusions:

• StriDe offers a powerful new tool for genome assembly.
• The hybrid approach effectively addresses limitations of existing methods.
• StriDe provides competitive performance in both contiguity and accuracy.