This article seems to give a good review: http://www.annualreviews.org/doi/pdf/10.1146/annurev.micro.55.1.709

We can distinguish two general principles for detection:

• Based on sequence composition alone (e.g. atypical codon usage or base compositions)
• Based on sequence comparison (atypical homology with genes in other species, phylogenetic methods: e.g. comparison of phylogenetic trees of different gene sets)

In principle it seems that such methods work best on the DNA sequences of the genomes in question, if you don't have the sequences, you might have to sequence the genomes or at least some phylogenetic markers on DNA level.