Supplementary information to ‘Relating underrepresented genomic DNA patterns and tiRNAs: the rule behind the observation and beyond’

Technical details of the statistical model and the characteristics of the underrepresented patterns. The frequency profile approach Here we describe and analyse the DNA sequences in terms of a hierarchical statistical model (see Fig. S1 for graphic representation). Our motif is a head and tail pair of fragments separated by a variable length spacer. Considering the length of the units of our model, the size of the complete genome and the restricted genomic alphabet of the 4 bases, the head or the tail alone can not be sufficiently specific for any gene regulatory purpose as the number of model units is only 2560. However, in the case of functionally linked head and tail the number of variants is 2560 by 2560 ~ 6.5 million providing more than 3 orders of magnitude better specificity. In our model ‘link ’ would mean that a particular head and tail pair with a critical spacer between them fulfils a specific biological functionality, but the same head-tail couple is not effective at any other relative sequential positions