HIV linkage paper rejected

The initial linkage paper on HIV has been rejected from BMC retrovirology. This doesn't really surprise me since I really didn't think a biology paper would publish a purely computational analysis.

There were two main comments shared by all three reviewers:

1. Predicted linkages may be due to ancestral evolution instead of biological relevance. Our computational method has no way to distinguish between the two.
2. Biologists really don't care about the sequence linkages between two distant regions of the genome "At present the manuscript describes a computational method that is not (yet) of interest to retrovirologists and should be submitted to a computational journal."

The reviewers quoted a few other potential sources that would account for the linkage we observe:

• Ancestral Origin (founder effects)
• RNA motifs
• CTL selection

I agree that all of these are possible choices ... I would just argue that I really didn't care what made the linkages I only care that given a small section of the genome I can predict a large portion of the rest of it. I think when we started adding too many biological examples this message got watered down and it seemed that we were claiming a biological reason.

I'm going to start brainstorming how I can deal with the phylogenetic biasing ... since I agree that this may be inflating my "non-perfect" scores ... ie if one node of my tree is overrepresented and it has a strong linkage then it could inflate my scores. In the perfect matches it doesn't actually matter how I slice the data, I will always get a perfect score.