Definition of the transcriptional units of inherited retinal disease genes by meta-analysis of human retinal transcriptome data

Abstract

Background: Inherited retinal diseases (IRD) are genetically heterogeneous disorders that cause the dysfunction or loss of photoreceptor cells and ultimately lead to blindness. To date, next-generation sequencing procedures fail to detect pathogenic sequence variants in coding regions of known IRD disease genes in about 30-40% of patients. One of the possible explanations for this missing heritability is the presence of yet unidentified transcripts of known IRD genes. Here, we aimed to define the transcript composition of IRD genes in the human retina by a meta-analysis of publicly available RNA-seq datasets using an ad-hoc designed pipeline.

Results: We analysed 218 IRD genes and identified 5,054 transcripts, 3,367 of which were not previously reported. We assessed their putative expression levels and focused our attention on 435 transcripts predicted to account for at least 5% of the expression of the corresponding gene. We looked at the possible impact of the newly identified transcripts at the protein level and experimentally validated a subset of them.

Conclusions: This study provides an unprecedented, detailed overview of the complexity of the human retinal transcriptome that can be instrumental in contributing to the resolution of some cases of missing heritability in IRD patients.

Keywords: Alternative splicing; Human retina; Inherited retinal disease; RNA-seq; Transcriptome.

Fig. 1

Analysis of IRD gene transcriptional units. A Number of already annotated and novel IRD transcripts with a scaled median TPM > 1 belonging to the analysed IRD genes. B Distribution of IRD genes according to their number of total distinct transcripts. C Distribution of IRD genes according to their number of already annotated transcripts (orange bars) and newly predicted transcripts (blue bars). D Correlation between the total number of distinct transcripts per gene and the genomic span of the corresponding gene. X-axis, number of IRD genes; Y-axis, genomic span (bp) in log2

Fig. 2

Novel transcripts description. A Schematic representation of different types of newly identified transcripts, which result from alternative splicing events that lead to intron retentions (i), exon extensions (ii), novel exon inclusion (iii), exon skipping (iv) and exon shortening (v). Example of a connection between two transcriptional units is represented by isoform (vi) that connects Gene 1 and 2. B Relative contribution to overall gene expression by annotated and novel IRD transcripts. X axis, relative contribution (in %) of gene expression in bins of 10, Y axis, number of IRD genes. C Structural classification of novel HEITs (highly expressed IRD transcripts). New exons are the most predominant feature, followed by intron retentions, while exon shortenings are the least common events

Fig. 3

RT-PCR validation of a subset of HEITs. Left panels: schematic representations of each analysed transcript (depicted in blue) with respect to its corresponding canonical form, as defined in the Ensembl database (depicted in orange). Forward and Reverse oligonucleotide primers used in RT-PCR experiments to validate the distinguishing features of the novel transcripts are depicted as red arrows in the higher magnification insets (circles). Right panels: Agarose gel electrophoresis of RT-PCR products. A Validation of a PLA2G5 transcript containing a novel exon. B Validation of a GRK1 transcript containing a novel exon. C Validation of a PRPH2 transcript containing a novel exon skipping event. The RT-PCR product corresponding to the canonical PRPH2 transcript is indicated in orange, whereas the product corresponding to the novel isoform is in blue. D Validation of a KIAA1549 transcript that harbours a novel exon. E Validation of a MAK transcript that contains an alternative last coding exon). F Validation of an RDH5 transcript that is connected with the adjacent BLOC1S1 transcriptional unit. G Validation of a MERTK transcript that contains a novel exon. Please note that the images showing the RT-PCR results of GRK1 (B) and KIAA1549 (D) were cropped and reorganized for the sake of clarity (source data are shown in Supplementary Fig. S2). L, 1000 bp ladder; B, Blank; BL, Blood; R, Retina; R1, Retina1; R2, Retina2; PO, Podocytes; F, Fibroblasts