New de novo assembly of the Atlantic bottlenose dolphin (Tursiops truncatus) improves genome completeness and provides haplotype phasing

Abstract

High-quality genomes are essential to resolve challenges in breeding, comparative biology, medicine, and conservation planning. New library preparation techniques along with better assembly algorithms result in continued improvements in assemblies for non-model organisms, moving them toward reference-quality genomes. We report on the latest genome assembly of the Atlantic bottlenose dolphin, leveraging Illumina sequencing data coupled with a combination of several library preparation techniques. These include Linked-Reads (Chromium, 10x Genomics), mate pairs (MP), long insert paired ends, and standard paired end. Data were assembled with the commercial DeNovoMAGIC assembly software, resulting in two assemblies, a traditional "haploid" assembly (Tur_tru_Illumina_hap_v1) that is a mosaic of the two parental haplotypes and a phased assembly (Tur_tru_Illumina_phased_v1) where each scaffold has sequence from a single homologous chromosome. We show that Tur_tru_Illumina_hap_v1 is more complete and more accurate compared to the current best reference based on the amount and composition of sequence, the consistency of the MP alignments to the assembled scaffolds, and on the analysis of conserved single-copy mammalian orthologs. The phased de novo assembly Tur_tru_Illumina_phased_v1 is the first publicly available for this species and provides the community with novel and accurate ways to explore the heterozygous nature of the dolphin genome.

Keywords: Tursiops truncatus; de novo genome assembly; 10x Genomics; DeNovoMAGIC; Illumina; bottlenose dolphin.

Figure 1:

Venn diagram of BUSCOs present in two dolphin assemblies. Out of 4,104 BUSCOs in the mammalia set, 105 are missing from both assemblies. Our assembly has 165 BUSCOs not present in Tur_tru v1, and Tur_tru v1 has 55 BUSCOs that are not present in our assembly.

Figure 2:

An example mummerplot of the alignments of the phased assembly to the “haploid” one, spanning about 15 Mbp of sequence of scaffold314. The circles represent contig ends, with lines joining them representing aligned sequence. The color indicates the direction of the alignment, with red and blue forward and reverse, respectively. We show that for most locations on the x-axis (haploid assembly coordinates) there are two alignments on the y-axis corresponding to the two phased haplotypes. The small number of regions with a single contig aligning represent long homozygous regions of the genome that we were unable to phase.

Figure 3:

This figure shows the alignment of the Tur_tru_Illumina_hap_v1 assembly to the human GRCH38 reference (primary chromosomes only). Each dot represents an alignment, with red indicating forward direction and blue indicating reverse direction. Human reference coordinates are on the x-axis and Tur_tru_Illumina_hap_v1 assembly alignment coordinates are on the y-axis. (A) Alignment of the entire assembly to the human reference with alignments to human chromosome 1 highlighted by the black box. One can clearly see the synteny that is present between the dolphin scaffolds and human chromosome 1. No other human chromosome shows clear synteny. Dolphin scaffolds with syntenic alignments spanning over 50% of the scaffold were extracted. (B)Alignments only to human chromosome 1.