Fig. 4: Synthetic MPRA variants lead to in vivo change of function in transgenic assay. | Nature Communications

Fig. 4: Synthetic MPRA variants lead to in vivo change of function in transgenic assay.

From: Massively parallel reporter assays and mouse transgenic assays provide correlated and complementary information about neuronal enhancer activity

Fig. 4

a Every fourth nucleotide of five MPRA tiles was mutagenized individually, for a total of 67 mutagenized constructs per reference tile. Dots connected by a vertical line = three biological replicates. Red horizontal line = zero, mean activity of scramble negative controls. Black horizontal line = mean activity of the reference construct. b Constructs encompassing the MPRA tiles, with or without the variants indicated in orange in panel A, were tested for activity using transgenic assay in developing mouse embryos at embryonic day e11.5. All variants except hs268.1 led to change of function in one or more neural tissues - brain, neural tube or cranial nerves. Shown are embryos that were genotyped as “tandem”, i.e. positive for insertion at the safe harbor locus and presence of the plasmid backbone indicating multi-copy insertion with strong, reproducible pattern. White arrowhead indicates loss of function, black indicates gain of function. See Supplementary Fig. 6 for all embryo images, including those genotyped as “tandem” and “single” (see Methods), which provide additional support of the changes observed. c An example of TFBS predicted by motifbreakR to be affected by the in vivo tested variants in hs978.1. Left: TFBS prediction consistent with all MPRA variant effects (POU4F3), right: TFBS prediction inconsistent with MPRA effects (CDX1 gain). Predicted TFBS match change and measured MPRA activity change symbols are colored green if matching and red if not. Arrowheads indicate an increase or decrease, flat line indicates no effect (assuming predicted TFBSs are activating). d Prediction of TFBSs likely affected by the variants and partially or fully consistent with flanking variant effects. Asterisk - TFBS events that are predicted to be created by in vivo tested variants are assumed to be unaffected by loss-of-activity flanking variants i.e. no loss of binding is possible, where no binding was observed in the reference sequence.

Back to article page