Fast and efficient Rmap assembly using the Bi-labelled de Bruijn graph.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Genome-wide optical maps provide high-resolution restriction maps of genomes.
• Current Rmap assembly methods are limited, with only one non-proprietary OLC-based method that struggles with large genomes.
• Proprietary software exists but its algorithm is largely unknown.

Purpose of the Study:

• To develop a novel, scalable Rmap assembly method using de Bruijn graphs.
• To introduce RMAPPER, the first de Bruijn graph-based assembler for optical mapping data.
• To compare RMAPPER's performance against existing methods.

Main Methods:

• Extended the definition of bi-labels in paired de Bruijn graphs for optical mapping data.
• Implemented RMAPPER, a de Bruijn graph-based Rmap assembler in C++.
• Benchmarked RMAPPER against Valouev et al.'s method and Bionano Solve on E. coli, human, and Anabas Testudineus genomes.

Main Results:

• RMAPPER successfully assembled all tested genomes, unlike Valouev et al.'s method which only worked on E. coli.
• On the human genome, RMAPPER was 130x faster and used 5x less memory than Bionano Solve.
• RMAPPER achieved the highest genome fraction with zero mis-assemblies.

Conclusions:

• RMAPPER provides a scalable and efficient solution for genome-wide optical map assembly.
• The de Bruijn graph approach overcomes limitations of OLC methods for large genomes.
• RMAPPER is publicly available, promoting open-source development in bioinformatics.