FIGURE 9.

The hexamer composition of predicted lncRNAs described in large-scale transcriptome analyses resembles that of 3′ UTRs and is markedly different from 5′ UTRs and ORFs of protein-coding genes. (A) The overlap between the three sets of hexamers that show a fivefold over- or under-representation in each of the three groups of noncoding sequences compared with the 5′UTR+ORF of protein-coding RNAs. The low number of hexamers in the functional lncRNAs class is due to the smaller number of the RNAs in this group. (B) The hexamer composition of predicted lncRNAs, 3′ UTRs, and 5′ UTRs and ORFs. The analysis is done in an identical fashion to that in Figure 7, but instead of functionally characterized lncRNAs, the lncRNAs described in a large-scale transcriptome study have been analyzed (Guttman et al. 2009; Khalil et al. 2009). The hexamers that showed more than fivefold difference in representation between any two of the three groups of sequences are shown in the figure.