“…The profile of ribonucleotides adjacent to deletion sites is unlike what is seen at insertion sites+ As a point of reference, the baseline percentage of ribonucleotides adjacent to deletion sites is lower because there are many fewer deletion sites than insertion sites in edited regions of RNA: 5+79% of ribonucleotides in edited regions are located adjacent to a deleted U+ The proportions of As, Gs, and Cs upstream of inserted Us vary less than at insertion sites, and are not significantly different from random (see Fig+ 3B)+ Us adjacent to deletion sites, consistent with the current mechanistic model of U deletion (Hajduk et al+, 1997;Stuart et al+, 1997;Estevez & Simpson, 1999), are considered to be downstream of deleted Us+ This is supported by studies of in vitro U deletion, which indicate that all Us 59 to the position of cleavage at the editing site are removed down to the first non-U ribonucleotide, and thus, that nondeleted Us are not upstream of deleted Us (CruzReyes & Solner-Webb, 1996; Seiwert et al+, 1996; Lawson et al+, 2000)+ The relatively high proportion of Cs upstream of deletion sites contrasts with the strong bias against C upstream of insertion sites+ The finding that Cs are well represented in this position at deletion sites is surprising+ This context was suggested to pose a particular difficulty because the complimentary G in the gRNA would be predicted to pair with the U to be deleted (Feagin, 1990;Sturm et al+, 1992)+ This was termed the "CU paradox" (Sturm et al+, 1992), and was suggested to possibly be a contributing factor to misediting at deletion sites+ However, the relatively favorable probability that a C in an edited region will be found in this context suggests that it is not selected against, and raises the pos- The same analysis performed for ribonucleotides upstream of deleted Us+ For this analysis, encoded Us are considered never to be upstream of deleted Us+ None of the percentages are statistically different from random at the level of 95% confidence: p-values for A, G, and C are 0+207, 0+052, and 0+128, respectively+ sibility that the extension of the anchor duplex by a G:U pair does not negatively influence U deletion at that site+ A relatively high proportion of encoded Us in edited regions are located downstream of deletion sites (Fig+ 3A)+ The fraction of deletion sites with a G or a C downstream is nearly what would be expected by chance+ As are present in this position slightly less often than would be expected by a random distribution+ This is unlike the distribution of ribonucleotides downstream of insertion sites, at which As and Gs predominate+ Both upstream and downstream of insertion sites, the ribonucleotides that are found more often than would be expected by chance are those that imply complimentary gRNA bases that do not pair with U+ The profile of ribonucleotides adjacent to deletion sites indicates no such bias+…”