The plots are quite striking. Below are two: one gene for which mutants survive selection in the lung (hemR), and one that's required (rfaF).
I computed statistics in order to recreate Fig 2 from the paper, but there's something wrong with my code. At least, my plot looks different because I am not getting the same values for the survival index as they did. Rather than post something that's probably incorrect, I won't show that. I'm suspicious that the BLAST step is an issue, because there are a number of assignments to positions that are not TA dinucleotides, although the positions with larger numbers of observed hits do tend to be of this type.
But I will show the results filtered for likely genes involved in survival or growth in the lung. If you know anything about bacterial physiology (or read the paper), the suspects on the list will not surprise you. Note particularly dam, ftsEKX, hel, polA, recABC, ruvABC and xerD, genes needed for LPS like galU, galE, and rfaF, and genes involved in amino acid and purine synthesis such as (aspC, guAB, ilvE, metC, purB, serAB, thiL, trpAC), and finally, Mike Cashel's favorite, relA. But probably the really interesting ones are the genes with "HI" designations.
I still want to look at tools like BLAST, BLAT, SOAP, uclust and bowtie for the genome position assignment step. That'll have to be some other time.