Summary of the deep sequencing analysis of low-molecular-weight RNAs and the extraction of novel sRNAs in E. coli. (A) Pie charts classifying the deep sequencing reads. [a] The percentages of all the deep sequencing reads mapped to the previously annotated or non-annotated regions of the E. coli genome. [b] The relative proportions of the deep sequencing reads (n = 3,022,146) mapped to the annotated regions. 'Other' includes the reads that mapped to pseudogenes or phantom genes. The reads that mapped to several genes of various types were also classified as 'Other'. [c] The relative percentages of the deep sequencing reads (n = 43,060) mapped to the non-annotated regions. IGR, intergenic region; Antisense [RBS], reads located on the cis-antisense strand of the ribosomal binding site of a known gene; Antisense [other], reads located on the cis-antisense strand of a known gene; Other, reads mapped to both intergenic and cis-antisense regions. (B) Schematic representation of the procedure used to extract and classify novel candidate sRNAs. Novel candidate sRNAs were extracted from all novel transcribed regions by the following two steps: selection of transcribed regions with evidence of transcription initiation (Step 1) and length (Step 2). Then, novel candidate sRNAs were classified according to their evidence for transcription initiation and their coding positions in the E. coli genome; Class A, with both computationally predicted sigma 70 promoters and experimentally determined TSSs [30–33] and/or RBRs ; Class B, with only computationally predicted sigma 70 promoters; Class C, with only experimentally determined TSSs and/or RBRs.