SAGE detects microRNA precursors

Analysis of 29 human and 230 mouse longSAGE libraries revealed the expression of 22 known and 10 predicted mammalian miRNAs. Most were detected in embryonic tissues. Four SAGE tags detected in human embryonic stem cells specifically match a cluster of four human miRNAs (mir-302a, b, c&d) known to be expressed in embryonic stem cells. LongSAGE data also suggest the existence of a mouse homolog of human and rat mir-493.The observation that some orphan longSAGE tags uniquely match miRNA precursors provides information about the expression of some known and predicted miRNAs.MicroRNAs (miRNAs) are endogenous, ~22 nucleotide (nt) noncoding RNAs that play important roles in gene expression regulation by base-pairing with messenger RNAs [1]. A single miRNA can down-regulate a large number of target mRNAs [2]. Since most miRNA precursors can be mapped to ~60–120 nt long conserved genomic regions and can be folded into hairpin structures, miRNAs can be predicted from genomic sequences with high sensitivity [3-9]. Experimental confirmation and functional analysis of these predicted miRNAs, however, remains a challenge.Serial analysis of gene expression (SAGE) collects short 14–21 nt tags from 3' ends of transcripts after certain restriction enzyme cutting sites; the most frequently used site is "CATG" which is recognized by NalIII [10] recently developed variation of this technique known as longSAGE collects 21 bp tags, which are long enough for genomic mapping and specific annotation [11]. Unlike DNA microarray that depends on a pre-defined gene set, SAGE is an exploratory method for transcriptome analysis. Many orphan SAGE tags that cannot be associated with any known transcripts represent potential novel transcripts [12].Primary miRNAs transcribed by polymerase II are processed by the nuclear Drosha enzyme to give pre-miRNAs, which are then exported into cytoplasm and lead to mature miRNAs. At least some primary miRNAs are known to be capped and polyadenylated in the nucleu