Figure 1.
Selfish and Cooperative Transposition, and the Game ISs Play
(A) shows the basic structure of a composite transposon. The narrow open bar indicates the selectable gene(s). The wide bar inscribed with ISL and ISR indicates the ISs flanking these genes. Gray triangles indicate the inverted repeats that constitute the inside and outside ends of the ISs. Red triangles indicate repeat sequences used in a particular type of transposition event. The upper panel illustrates selfish transposition, where either IS transposes independently from the other. The lower panel illustrates cooperative transposition, where the composite transposes as a whole, taking the selectable genes with it.
(B) The structure of the payoff matrix P describing the payoff to ISL in the symmetric game, where both ISs have the same target specificity and the same likelihood of undergoing DNA rearrangement. The parameter t is the joint probability of transposition and horizontal transfer of the transposition product, r is the probability of DNA rearrangement associated with cooperative transposition, and s is the probability that the new host of a transposable element will need the selectable genes for its survival. Entry PSC is the payoff to ISL if a selfish ISL is paired with a cooperative ISR in a composite, and analogously for the other entries of P. The payoff matrix for ISR is the transpose of P.
(C) Schematic diagram of the evolutionary stable strategies (ESSs) in the case of genetic dominance of selfish transposition. Selfish transposition (“S”) is the only ESS if r > ts. Selfish and cooperative transposition (“C”) are both ESSs if r < ts.
Figure 2.
No Significant Enrichment of Composite Transposons in 376 Completely Sequenced Bacterial Genomes and Plasmids
(A) The horizontal axis shows the number of nearest-neighbor IS pairs at a distance between 0.5 and 30 kb in the bacterial genomes examined. The p-value on the vertical axis is the likelihood that this number of IS pairs is greater than expected by chance alone, according to a randomization test described in Materials and Methods. There are two horizontal lines near the horizontal axis. The upper line represents a p-value of 0.01, the lower line a p-value of 0.00064, which is the Bonferroni-corrected value of p = 0.05, given that n = 78 independent statistical tests carried out for this analysis (0.00064 = 0.05/78).
(B) The randomized distribution of nearest-neighbor distance for IS91 in the genome of Rhodopirellula baltica. There are five members of this IS in the genome, and the closest pair of them is 3,051 bp apart. Based on this randomization test, this distance is significantly closer than expected by chance at a p = 0.0074.
(C) The horizontal axis shows the minimal distance between nearest-neighbor IS pairs that are between 0.5 and 30 kb apart in a genome. The vertical axis shows the likelihood that this minimal distance is observed by chance alone, based on the randomization test from (B) (described in Materials and Methods). The upper and lower horizontal lines represent Bonferroni-corrected p-values of p = 0.05 (0.00064 = 0.05/78) and p = 0.01 (0.000128 = 0.01/78). Both panels contain n = 78 datapoints derived from ISs in 28 different families.