The ribozyme approach distinguishes RNA-dependent and RNA-independent functions of lincRNA genes. The ribozyme approach distinguishes RNA-dependent and RNA-independent functions of lincRNA genes. (A) Diversity of knockdown efficiencies for 15 selected lincRNAs, targeted by genomic integration of the HHRz and assessed using RNA-seq analysis across the collection of 140 cell lines. The expression of the lincRNAs targeted in the knockdown cell lines is shown, normalised to the median of all other cell lines. Each point corresponds to a biological replicate/cell line and the bar is the median across replicates. G9a mRNA expression across the G9a HHRz cell lines is shown for reference. (B) Representative knockdown of linc1509 using the HHRz. RNA-seq reads across the linc1509 pre-lincRNA (including introns) are shown, comparing the average read depth in non-targeted cell lines (black) with that in cell lines with the HHRz in linc1509 exon 2 (top) or exon 3 (bottom) (red). (C) Rescue of linc1405 expression in HHRz depleted linc1405 cell lines using an oligonucleotide that blocks ribozyme cleavage. Linc1405 expression measured by qPCR across a wild-type cell line (white) and a linc1405 HHRz cell line (red), treated with increasing concentrations of the oligonucleotide. As a negative control, the linc1405 HHRz cell line was treated with a non-complimentary oligonucleotide (“−ve control”). Dots represent individual measurements (two biological replicates, each with three technical replicates), and statistical significance was assessed using a two-tailed t test. (D) Expression of genes neighbouring eight lincRNA loci, in cell lines where those lincRNA loci were targeted by HHRz integration (coloured triangles) or genomic deletion (coloured crosses). Gene expression was measured by RNA-seq and normalised to median expression in non-targeted cell lines (grey points). Median expression differences between targeted and non-targeted groups of cell lines were tested for significance using a Monte Carlo simulation, applying the Bonferroni correction for multiple hypothesis testing. (E) Expression of linc1405 (red) and the neighbouring gene Eomes (black) across the differentiation time course, measured by bulk RNA-seq (n = 2 biological replicates). For each gene, measurements are centred on the mean and scaled by standard deviation. (F) Assessment of correlated expression across mESCs and NPCs of cis-linked gene pairs involving one lincRNA gene, quantified by Euclidean distance using single-cell RNA-seq data. The % of gene pairs (in 200 kb bins) with Euclidean distances below the 1% significance threshold (determined in Fig S4A by examining distant gene pairs) is shown. In the absence of proximity-dependent effects, we expect only 1% of gene pairs to be correlated below this threshold. (G) Representative single-cell RNA-seq tracks (three mESCs and three NPCs) for the genomic neighbourhood around linc_2485 (red bar). Genes co-expressed with linc_2485 (according to Fig S4A) are indicated as black bars. See also Figs S3 and S4 and Table S4A and B. Alex C Tuck et al. LSA 2018;1:e201800124 © 2018 Tuck et al.