I agree with Artur.  If you were sequencing a protein-coding gene, you could look for stop codons or compare the synonymous to nonsynonymous substitution ratio (in a functional gene, you would expect all the changes to be synonymous among closely-related taxa, whereas in a pseudogene, all sites are equally likely to exhibit changes).  However, the D-loop does not provide these tools.

Thank you so much (Arthus and Andrew) for your suggestions and comments.

One thing you can still look for is the presence of heterozygous sites. If you have heterozygous sites in an individual then you know you have a NUMT. Other wise since this is non-coding there is no real way to tell.

Best Regards

I am also assuming you do not know whether you have NUMTs or not? If you know that you do, then as Artur has said, just dump them unfortunately.

Heterozygous sites do not necessarily indicate NUMTS since heteroplasmy is also a possibility.

briefly discussed in:

Galtier et al. 2009. Mitochondrial DNA as a marker of molecular diversity: a reappraisal. Molecular Ecology 18(22):4541-4550.

The horse genome is full of NUMTs from every part of the mtDNA including CR. If your sequences match to any of these that could help indicate whether you have a true mtDNA copy or not.

Nergadze et al. 2010. Mitochondrial DNA insertions in the nuclear horse genome. Animal genetics 41(s2):176-185.

Alternatively, if there is still tissue for some of the specimens you used you could re-extract the mtDNA only (e.g. using ultracentrifugation), re PCR and compare the sequences to those you originally made:

Triant DA, deWoody JA. 2007. The occurrence, detection, and avoidance of mitochondrial DNA translocations in mammalian systematics and phylogeography. Journal of Mammalogy 88(4):908-920.

Or either follow the post extraction advice given in:

 Calvignac S, Konecny L, Malard F, Douady CJ. 2011. Preventing the pollution of mitochondrial datasets with nuclear mitochondrial paralogs (numts). Mitochondrion 11(2):246-254.

and compare the sequences with those you originally generated.

Good luck!

Dear Serena,  

You are right. I don't know if the sequences we obtained contains NUMTS since we overlooked the procedure to exclude NUMTS during amplification. I will try out the options provided and see it if solves the problem. 

Thank you once again.