Table 2. Mutual confirmation of novel transcriptional elements in Geuvadis and Encode RNA-Seq data.

	Geuvadis	Encode	Novel
(a) PNIs
All	1,068,786	62,5%	400,697
>0 in all populations	205,649	94,9%	10,553
>150 individuals	21,761	97,8%	469
>300 individuals	6,660	97,9%	139
>450 individuals	846	97,4%	22
(b) PCSs
all PCSs ≥2 mappings	21,102	62,8%	7,856
not overlapping 3′ UTR	6,032	40,2%	3,607
overlapping 3′ UTR	15,070	86,6%	2,017

The table presents the number of different (subsets of) novel transcriptional elements (rows) predicted from the Geuvadis experiments (column 2), the proportion of these elements that is additionally confirmed by Encode RNA-Seq reads (column 3), and the number of non-overlapping (i.e., novel) elements in Geuvadis as compared to Encode. (a) Nearly 2/3 (~63%) of the putative novel introns (PNIs) in Geuvadis are also contained in the 34,926,167 Encode PNIs. Applying more restrictive population support thresholds on the PNIs leads to high confirmation rates (>97%). (b) Putative cleavage sites (PCSs) with a support of ≥2 RNA-Seq reads show similar a base level (~63%) of overlap between the Geuvadis data set and the 160,331 PCSs correspondingly rescued from the Encode data. For PCSs outside annotated 3′ UTRs the confirmation rate decreases (~40%), whereas PCSs in 3′ UTR regions are strongly supported by Encode data (~87%).