An attempt at a higher-level phylogeny of the ants appeared in the most recent Molecular Phylogenetics and Evolution:
Astruc, C., J. F. Julien, C. Errard and A. Lenoir. 2004. Phylogeny of ants (Formicidae) based on morphology and DNA sequence data. Molecular Phylogenetics and Evolution 31:880-893.
In order to reconstruct ants' phylogeny, we analysed DNA sequences for two nuclear genes, abdominal-A and Ultrabithorax, from 49 species of ants and two outgroups. As these genes control the development of the first segments of the abdomen in insects, which are very variable in ants (petiole, postpetiole, and gaster constriction), we hypothesized that the morphological variations between the subfamilies may be correlated with mutations of some abd-A or Ubx regions. Contrarily to our hypothesis, these sequences are highly conserved. The differences observed concern mainly third codon positions and present some saturation. Phylogenetic reconstructions were carried out using the genetic raw sequence data and by combining them with a set of morphological data (Total Evidence). Relations among subfamilies of ants remains poorly resolved with molecular data only, but adding these data to morphological characters confirms and reinforce the topology of Baroni Urbani et al. (1992): a Poneroid complex [Ponerinae, Cerapachyinae, Leptanillinae and army ants], a Formicoid complex [Dolichoderinae, Formicinae] and a Myrmecoid complex [Myrmicinae, Myrmeciinae, Pseudomyrmecinae, Nothomyrmeciinae]. Our molecular results allow resolution near the branch tips and three subfamilies (Dolichoderinae, Formicinae and Pseudomyrmecinae) always appear as monophyletic. The Formicinae and the Dolichoderinae have close relationships. The Camponotini appear as a strong clade inside the Formicinae. The Ponerinae are separated in two parts: the Ectatommini and all other tribes. The Cerapachyinae, Dorylinae, and Ecitoninae belong to the same clade, the Cerapachyinae being confirmed in their subfamily status. The Myrmicinae appears to be very heterogeneous, with the Attini forming a very stable and well-separated group.
There is, to be honest, less here than meets the eye. Astruc et al give a valient effort at a thorough taxon sampling but the genes they employed- Abd-A and Ubx- do not give much signal at the more interesting deeper levels. The resolution in the combined morphological and molecular analysis seems to be driven largely by morphology, and the results consequently provide little new insight into deep ant history. However, the question of formicid phylogeny remains an interesting one, and as new genes become available for phylogenetic analysis we should expect some real progress on this front in the not-too-distant future.