MitoZ: a toolkit for animal mitochondrial genome assembly, annotation and visualization

Abstract

Mitochondrial genome (mitogenome) plays important roles in evolutionary and ecological studies. It becomes routine to utilize multiple genes on mitogenome or the entire mitogenomes to investigate phylogeny and biodiversity of focal groups with the onset of High Throughput Sequencing (HTS) technologies. We developed a mitogenome toolkit MitoZ, consisting of independent modules of de novo assembly, findMitoScaf (find Mitochondrial Scaffolds), annotation and visualization, that can generate mitogenome assembly together with annotation and visualization results from HTS raw reads. We evaluated its performance using a total of 50 samples of which mitogenomes are publicly available. The results showed that MitoZ can recover more full-length mitogenomes with higher accuracy compared to the other available mitogenome assemblers. Overall, MitoZ provides a one-click solution to construct the annotated mitogenome from HTS raw data and will facilitate large scale ecological and evolutionary studies. MitoZ is free open source software distributed under GPLv3 license and available at https://github.com/linzhi2013/MitoZ.

Figure 1.

MitoZ toolkit components. Ellipses indicate input data files, rectangles (solid line) represent functional modules in MitoZ which can be run independently when users provide corresponding input files.

Figure 2.

MitoZ test result. (I) the proportion of assembled mitogenomes in different assembly categories; (II) diagram of different assembly types where the boxes represent PCGs or rRNA genes and the solid lines stand for the other parts of mitogenomes and the upper grey boxes represent sanger mitogenome while the lower colorful ones are assembled by MitoZ; (III) the gene completeness distribution of PCGs and rRNA genes of the 50 test species. ‘full length’ means gene completeness ≥ 95%, ‘partial’ means gene completeness < 95%, and ‘missing’ means the genes that were not recovered by MitoZ.

Figure 3.

Gene similarities (MitoZ versus Sanger). The blue dots present three species whose genes possessed similarities < 97% to their sanger mitogenomes but can find high similarity genes of the same species in NCBI NT database. Such incongruences could derive from intraspecies variances. The red points (five in total) present genes possessed similarities < 97% to their sanger mitogenomes and could not find better hits in NCBI. The rest genes and samples were presented by green dots.

Figure 4.

Demonstration of mitogenome visualization using MitoZ.

Figure 5.

Performance comparisons between MitoZ and NOVOPlasty. (I) Left: mitogenome types, see Figure 2(I) and Figure 2(II) for the categories of mitogenome types, while type D indicates the total number of PCG and rRNA genes recovered was less than eight. Right: gene (PCGs and rRNA genes) completeness distribution, see Figure 2(III) for the meanings of ‘full length’, ‘partial’ and ‘missing’. (II) Diagram of gene similarities to the sanger genes. Genes that recovered by both MitoZ and NOVOPlasty (n = 592) were included in the analysis, and paired samples t-test was used.