Abstract
The assembly of plant mitochondrial genomes presents unique challenges due to difficulties in isolating mitochondrial DNA (mtDNA) and plant mitochondrial genome characteristics, such as low interspecific conservation; sequence sharing among mitochondrial, nuclear and plastid DNAs; and complex structural variations. Our laboratory has sequenced and assembled a dozen plant mitochondrial genomes, testing various strategies and identifying numerous assembly issues. This review compared the advanced methods and tools for plant mitochondrial genome assembly, categorizing assembly algorithms into three groups: (1) reference-based, (2) de novo and (3) iterative mapping and extension. The performance of 11 software tools used most frequently over the past 5 years (GetOrganelle, Velvet, NOVOPlasty, SOAPdenovo2, Canu, Flye, SMARTdenovo, PMAT, NextDenovo, SPAdes and Unicycler) and two newly developed tools (TIPPo and Oatk) was assessed. The evaluation metrics included the completeness, contiguity and correctness of the assembled plant mitochondrial genomes. SMARTdenovo, NextDenovo and Oatk demonstrated superior performance in terms of contiguity and completeness. GetOrganelle and Flye excelled in correctness. Key challenges in plant mitochondrial genome assembly, such as removing nuclear mitochondrial DNA (NUMT) and mitochondrial plastid DNA (NUPT) contamination and resolving intra-genomic repetitive regions, were discussed. A general strategy for plant mitochondrial genome assembly used in studies conducted in our laboratory was summarized. This review serves as a resource for those assembling plant mitochondrial genomes or developing plant mitochondrial genome assembly tools.