Closing and Finishing Genomes with Long Span NGS Reads

Long range sequence linkage from short NGS reads are needed for accurate, economical de novo assembly of genomes. New library construction methods for producing long span NGS reads from 2-50Kb on the Illumina or Ion Torrent platform are described. User defined 2-8Kb mate pair (MP) libraries can be produced with bead based protocols. Gel purification enables user defined 10-25Kb MP libraries. Encrypted barcodes (chimera codes) are used to bioinformatically distinguish true MP from false junctions, resulting in high quality libraries (>95% true mates).  A novel fosmid vector has been used to construct 35-50 kb MP libraries. These methods have been extensively validated using an E. coli reference genome to assess closing and finishing efficiencies, and were subsequently used for de novo closing and finishing of multiple genomes. Without MP libraries the genome assemblies contained numerous, unordered contigs. A Thermus aquaticus genome (72% GC) was closed and finished using a combination of 8Kb MP and fragment libraries. A 2.5Gb Miscanthus sinensis genome (80% repeats) assembly was significantly improved with a 35-50Kb fosmid MP library. A 1Gb fish genome was assembled to a few thousand scaffolds with a combination of fragment, 8Kb and fosmid MP libraries. The addition of >95% efficient true MP data allows accurate de novo assembly and closing of microbial small genomes and a significant reduction in the number of scaffolds for complex, repeat rich genomes. The incorporation of long span NGS reads up to 50Kb is an effective way to order and orient contigs and close genomes.