C. elegans piRNAs are processed from capped-small RNAs transcribed at promoters throughout the genome

Piwi-interacting (pi) RNAs are a class of germline-expressed small RNAs that have been linked to epigenetic programming in metazoa. C. elegans piRNAs known as 21U-RNAs are defined by more than 15,000 genome-encoded species. To explore the origin of 21U-RNAs we employed methods to enrich the 5' ends of Pol II transcripts. We show that a species of capped-short (cs) RNA is frequently expressed bidirectionally at Pol II loci in C. elegans. Interestingly, at annotated 21U-RNA loci, csRNAs originate precisely 2 nt upstream of the mature piRNA species suggesting that csRNAs are piRNA precursors. In addition, we show that csRNAs associated with TS sites genome-wide define a previously overlooked class of 21U-RNA loci, and nearly double the number of piRNA species available for genome surveillance. Our methods should be of general utility in TS site identification and 5' anchored RNA-expression profiling. Identification of capped RNA including capped small RNA and long capped RNA in C. elegans. The mouse data are independent data to test the CapSeq sequencing protocol.