Table 1.

5′ and 3′ features of cis-NAT siRNAs

(a)				Read counts normalized per million reads
				21-mers in cis-NAT overlaps			Non–21-mers outside overlaps
Library	Reads	miRNAs	% miRNA	+ strand	− strand	Total	+ strand	− strand	Total
Dcr-2/CyO heads	908,403	833,686	91.8	50.6	53.9	104.6	12.1	7.7	19.8
Dcr-2/Dcr-2 heads	734,243	671,471	91.5	4.1	4.1	8.2	5.4	15.0	20.1

(b)		miRNA/hpRNA/siRNA counts normalized per 100,000 reads
Library	Total Reads	miRNAs	β↓	hpRNAs	β↑	Exo-siRNAs	β↑	Cis-NAT siRNAs	β↑
Head-β	42,354	16,128		1,354	20.0×	20,034 (white)	13.6×	92.0	7.0×
Head	105,976	35,746	2.2×	68		1,473 (white)		13.2
S2-β	113,709	3,920		811	5.8×	1,736 (GFP)	5.4×	137.1	4.9×
S2	99,003	31,785	8.1×	139		320 (GFP)		2,803

(a) cis-NAT siRNAs are dependent on Dcr-2. Previously published small RNA data²¹ was analyzed for miRNAs and cis-NAT RNAs. The miRNA content of Dcr-2 heterozygous and homozygous heads is similar, as is the content of non–siRNA-sized RNAs derived from non-overlap regions of cis-NAT pairs. However, there is a strong depletion of 21-mers derived from the overlap regions of 3′ cis-NAT pairs in Dcr-2 homozygous tissue. (b) Behavior of various small RNA classes following β-elimination. Relevant reads were extracted from adult head and S2 cell data, cloned before and after β-elimination²⁰, and normalized to counts per 100K. miRNAs are strongly depleted following β-elimination (β↓), whereas hpRNAs, white/GFP siRNAs and cis-NAT siRNAs are all strongly enriched (β↑).