Referee 3's Review:

**********
N.B. These are the comments made by the referee when reviewing an earlier version of this paper. Prior to publication, the manuscript has been revised in light of these comments and to address other editorial requirements.
**********

Overall, this paper is an excellent genome description/comparative analysis and is very well written. The genome is fascinating and the analysis sheds light on important questions about the genomic evolution in the myxobacteria and delta-proteobacteria more broadly. It was a pleasure to read and I strongly recommend it for publication after minor revision.

I find the primary evolutionary proposal of the paper, that this species derived from an aerobic ancestor, to be plausible and well presented with the exception of insufficient description of a large portion of the genome that is not most similar to other delta-proteos. See below.

Specific comments:

p2. What is the evidence that A. dehalogenans is prevalent? What is meant by this?
p3. I found some paragraphs to be too long for my own taste (also pp.5-6).
p4. M. xanthus is referred to as having a sequenced genome, but now the Sorangium genome is published as well.
p5. Perhaps the weakest point of the paper is that so little is said about the 50% of genes that are most similar to non delta-proteobacteria or the implications of this finding for the role of HGT in the evolution of this species away from its (likely) aerobic ancestor. This silence is particularly striking since the problem of the role of HGT in bacterial genome evolution is highlighted as a major motivating theme at the beginning of the paper.
Of the genes most similar to non-delta-proteobacteria, what proportion even have homologs in the M. xanthus or Stigmatella? Do any genes shared between A. dehalogenans and M. xanthus show greater similarity between the former and a non-delta-proteo species than with M. xanthus/Stigmatella?
p7. I think the statement about only 4% of the genome being attributable to HGT (line 157) needs qualification that that number refers only to HGT potentially detectable by the GC content criterion and is not an estimate how much of the genome is present due all HGT acquisitions since its common ancestor with M. xanthus, as the latter would seem highly inconsistent with the finding that half of all genes most closely match non-delta-proteobacteria.
p8. I think the scenario of genome evolution laid out here should reference and incorporate the authors' views of how the half of the genome most similar to non-proteobacteria fit into the story.
p10. I found the appeal to the r/K paradigm and classifying M. xanthus as an r species and A. dehalogenans as a K species to be somewhat questionable since so very little is known about the in situ ecology, growth rates and carrying capacities of these species outside the lab. M. xanthus grows very slowly compared to many other soil bacteria and its transition to and survival during stationary phase under non-lab conditions may differ greatly from that in the lab. Moreover, the value for the r vs. K framework itself has been questioned by multiple ecologists.
p11. It might be worth noting that Bdellovibrio also as TFP-type genes.
p12. The assumption that proximate genes have related functions seems asserted to strongly (e.g. "confirming the relationship", "implicating protein interactions") given the lack of relevant experimental data.
p14. The authors might check the Sorangium genome to see whether it contains any flagellar genes, which would further strengthen the case that the flagellar genes were present in the common myxo ancestor.
line 381. A reference to the source of the evolutionary gap hypothesis would be appropriate.