I agree with Robert Ekblom. The easiest, fastest and most reliable way to get complete mitochondrial genome sequences is in my opinion to extract total DNA from animal tissue and sequence a few million (5 to 10 million) 150 bp paired-end reads on an Illumina HiSeq sequencer (total cost 200 to 400 Euros per sample). 300 bp paired-end reads on the Illumina MiSeq would of course be an advantage. Once you have the reads you do not even need to do a reference guided assembly, just use an assembler optimized for bacterial de novo genome assembly (e.g. IDBA or SPAdes). Run an assembly with standard parameters and then pull the complete mitorchondrial genome from the assembled contigs for example by doing a blast search for the COI gene with a sequence of a related organism.

We have been doing a lot of metagenome sequencing of symbiotic animals recently, where we were actually interested in sequencing the bacterial symbionts. The mitochondrial genome was always a nice byproduct in these metagenomic analyses.

The first question is how much material you have? Depending on the amount you can use difeerent methods for obtaining mtDNA.

We used in the lab for many years: http://www.pnas.org/content/83/22/8649.full.pdf+html

You can combine it with different enrichment methods: differential centrifugation (as above), filtering thorugh nylon membranes, FACS,...

For sequencing, if you have good starting material, a good option should be PacBio.

Hi,

In my experience there is no need to specifically enrich for mtDNA. Just do a small amount of whole genome sequencing (for example on Illumina HiSeq or MiSeq). As there are many copies of mtDNA in each cell this will make up a substantial proportion (up to several percent) of the total sequence reads. Assemble the whole mtDNA by reference assisted assembly using a known mtDNA sequence from a related organism. For a recent paper where I did just this check out:

http://www.biomedcentral.com/1471-2164/15/467

Patterns of sequencing coverage bias revealed by ultra-deep sequencing of vertebrate mitochondria. BMC Genomics 2014, 15:467  

good luck!

/Robert

Article Patterns of sequencing coverage bias revealed by ultra-deep ...

I agree with Robert Ekblom. The easiest, fastest and most reliable way to get complete mitochondrial genome sequences is in my opinion to extract total DNA from animal tissue and sequence a few million (5 to 10 million) 150 bp paired-end reads on an Illumina HiSeq sequencer (total cost 200 to 400 Euros per sample). 300 bp paired-end reads on the Illumina MiSeq would of course be an advantage. Once you have the reads you do not even need to do a reference guided assembly, just use an assembler optimized for bacterial de novo genome assembly (e.g. IDBA or SPAdes). Run an assembly with standard parameters and then pull the complete mitorchondrial genome from the assembled contigs for example by doing a blast search for the COI gene with a sequence of a related organism.

We have been doing a lot of metagenome sequencing of symbiotic animals recently, where we were actually interested in sequencing the bacterial symbionts. The mitochondrial genome was always a nice byproduct in these metagenomic analyses.

As Robert and Manuel, I think that extracting total DNA and Illumina sequencing could work. See also Botero-Castro et al. 2013 for a recent example on Bats:

http://www.sciencedirect.com/science/article/pii/S1055790313002789

As mitochondria are in abundance the mitochondrial genome gets co-isolated along with nuclear genome.SO, sequencing the genomic DNA co-sequences the mitochondrial genome also. We have assembled mitochondrial genome from whole genomic reads of yeast, rice and Fusarium sp.