Long-read sequence and assembly of segmental duplications.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Genome assemblies often contain collapsed regions due to segmental duplications.
• Resolving these duplicated regions is crucial for accurate genome annotation and understanding genetic variation.

Purpose of the Study:

• To develop a computational method for resolving collapsed segmental duplications in genome assemblies.
• To improve the accuracy and completeness of genome assemblies by accurately assembling paralogous sequences.

Main Methods:

• Developed a computational method based on polyploid phasing of long sequence reads.
• Constructed graphs where paralogous sequence variants are nodes and long reads provide edges.
• Applied the method to single-molecule, real-time sequence data from three human genomes.

Main Results:

• Recovered 33-79 megabase pairs (Mb) of duplications in human genomes.
• Identified approximately 50% of loci diverged (<99.8%) from the reference genome.
• Demonstrated high accuracy (>99.9%) for the assembled sequences, revealing copy-number-variable paralogs absent from the reference.

Conclusions:

• The Segmental Duplication Assembler (SDA) method effectively resolves complex segmental duplications.
• The method can improve gene annotation and understanding of copy-number-variant genetic diversity in complex genomes.
• This approach aids in closing gene-rich gaps in genome assemblies.