Fig. 3

Correlation of effect of different IBD gene candidate ORFs on the HT-29 transcriptome. Correlation plots illustrating the shared effects of IBD gene candidate ORFs on the transcriptome of HT-29 cells are shown A for cluster 1 IBD causal genes IFIH1 vs NFKB1 and SBNO2, and NOD2 vs NFKB1, B for cluster 2 IBD causal gene KSR1 and its closest candidate ORF cluster neighbor DUSP16, as well as KSR1 causal gene HNF4A, and C for IBD causal gene SMAD3 and candidate ORF SMAD7. Additional correlation plots for IBD-causal genes with IBD gene candidate ORFs are shown in Additional file 2: Fig.S6. D Correlation plots comparing the effects of increased wild-type IFIH1 expression against expression of three of its IBD-associated non-synonymous coding variants: an isoleucine to valine substitution at position 923 (rs35667974) (923 V, left), a splice donor site variant at position + 1 in intron 8 (SP1, middle) and a splice donor site variant at position + 1 in intron 14 (SP2, right). For the effect of IFIH1 variants on IFIH1 HITs, see Additional file 1: Table S9. For all correlation plots shown, each dot represents a single detectable probe from the genome-wide array tagging a specific gene in the HT-29 transcriptome (see Fig. 1). The x-axis (inner color of dots) and y-axis (border color of dots) show the effect of two independent set of replicated ORFs on the transcriptome, as the log2-transformed fold-induction compared to baseline. Sky blue are probes with expression value within expected variation (|Z| ≤ 2), orange represent probes suggestively outside the range (|Z| > 2), red represent probes outside expected range of variation (|Z| > 4), and grey are probes with expression value below our detection threshold; this color code is applied to the inside of dots for x-axis data and the border of dots for y-axis data. No metric of correlation was included, as most probes are not affected by ORFS and thus only represent noise. See Fig. 2 for similarity illustration