摘要

Background: Problems associated with using draft genome assemblies are well documented and have become more pronounced with the use of short read data for de novo genome assembly. We set out to improve the draft genome assembly of the African cichlid fish, Metriaclima zebra, using a set of Pacific Biosciences SMRT sequencing reads corresponding to 16.5x coverage of the genome. Here we characterize the improvements that these long reads allowed us to make to the state-of-the-art draft genome previously assembled from short read data. Results: Our new assembly closed 68 % of the existing gaps and added 90.6Mbp of new non-gap sequence to the existing draft assembly of M. zebra. Comparison of the new assembly to the sequence of several bacterial artificial chromosome clones confirmed the accuracy of the new assembly. The closure of sequence gaps revealed thousands of new exons, allowing significant improvement in gene models. We corrected one known misassembly, and identified and fixed other likely misassemblies. 63.5 Mbp (70 %) of the new sequence was classified as repetitive and the new sequence allowed for the assembly of many more transposable elements. Conclusions: Our improvements to the M. zebra draft genome suggest that a reasonable investment in long reads could greatly improve many comparable vertebrate draft genome assemblies.

  • 出版日期2015-9-22