Abstract
The diversity of DNA sequencing methods and algorithms for genome assemblies presents scientists with a bewildering array of choices. Here, we construct and compare eight candidate assemblies combining overlapping shotgun read data, mate-pair and Chicago libraries and four different genome assemblers to produce a high-quality draft genome of the iconic coral reef Pearlscale Pygmy Angelfish, Centropyge vrolikii (family Pomacanthidae). The best candidate assembly combined all four data types and had a scaffold N50 127.5 times higher than the candidate assembly obtained from shotgun data only. Our best candidate assembly had a scaffold N50 of 8.97 Mb, contig N50 of 189,827, and 97.4% complete for BUSCO v2 (Actinopterygii set) and 95.6% complete for CEGMA matches. These contiguity and accuracy scores are higher than those of any other fish assembly released to date that did not apply linkage map information, including those based on more expensive long-read sequencing data. Our analysis of how different data types improve assembly quality will help others choose the most appropriate de novo genome sequencing strategy based on resources and target applications. Furthermore, the draft genome of the Pearlscale Pygmy angelfish will play an important role in future studies of coral reef fish evolution, diversity and conservation.