Konnector v2.0: pseudo-long reads from paired-end sequencing data

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Paired-end tag (PET) sequencing generates reads from both ends of DNA fragments.
• Gaps in sequencing occur when fragment length exceeds combined read length.
• Bioinformatics methods can bridge these gaps with sufficient sequencing coverage.

Purpose of the Study:

• To introduce Konnector v2.0, an improved local de novo assembly tool.
• To address the challenge of bridging gaps in PET sequencing data.

Main Methods:

• Konnector utilizes Bloom filters for memory-efficient k-mer spectrum representation.
• It constructs an implicit de Bruijn graph to identify k-mer overlaps.
• Graph traversal is employed to bridge sequencing gaps between read pairs.

Main Results:

• Konnector v2.0 demonstrates effective gap bridging on simulated and experimental datasets.
• The tool represents k-mers using an average of 1.5 bytes, enabling scalability to large genomes.
• A parallel implementation allows processing over a billion bases on standard hardware.

Conclusions:

• Konnector v2.0 offers a scalable and efficient solution for gap bridging in PET sequencing.
• Its performance is competitive with other tools of similar functionality.
• The tool facilitates comprehensive genomic analysis by enabling de novo assembly across gaps.