Multiplexed primer extension sequencing: A targeted RNA-seq method that enables high-precision quantitation of mRNA splicing isoforms and rare pre-mRNA splicing intermediates

Abstract

The study of pre-mRNA splicing has been greatly aided by the advent of RNA sequencing (RNA-seq), which enables the genome-wide detection of discrete splice isoforms. Quantification of these splice isoforms requires analysis of splicing informative sequencing reads, those that unambiguously map to a single splice isoform, including exon-intron spanning alignments corresponding to retained introns, as well as exon-exon junction spanning alignments corresponding to either canonically- or alternatively-spliced isoforms. Because most RNA-seq experiments are designed to produce sequencing alignments that uniformly cover the entirety of transcripts, only a comparatively small number of splicing informative alignments are generated for any given splice site, leading to a decreased ability to detect and/or robustly quantify many splice isoforms. To address this problem, we have recently described a method termed Multiplexed Primer Extension sequencing, or MPE-seq, which uses pools of reverse transcription primers to target sequencing to user selected loci. By targeting reverse transcription to pre-mRNA splice junctions, this approach enables a dramatic enrichment in the fraction of splicing informative alignments generated per splicing event, yielding an increase in both the precision with which splicing efficiency can be measured, and in the detection of splice isoforms including rare splicing intermediates. Here we provide a brief review of the shortcomings associated with RNA-seq that drove our development of MPE-seq, as well as a detailed protocol for implementation of MPE-seq.

Fig. 1.

Quantitation of splicing isoforms from RNA-seq data: (A) Only a subset of alignments are splicing informative: those which can be unambiguously mapped to premature (including completely unspliced and lariat intermediate) or mature (fully spliced) molecules. Black arrows denote the conversion between these forms in the 1^st and 2^nd chemical steps of the pre-mRNA splicing reaction. (B) The fraction of alignments from a standard RNA-seq experiment that are completely exonic (E), mature (J), and premature (EI + IE) is shown for all single intron containing genes in S. pombe. (C) The coefficient of variation of replicate measurements of each event in (B) is plotted as a function of the (geometric) mean count depth for the isoform. Count data for (B) and (C) can be found in Table A.1. RNA-seq data are from Xu, etal. (2019) [11].

Fig. 2.

MPE-seq targets reads to splicing informative loci, increasing the precision of measurements of splicing efficiency compared to standard RNA-seq. (A) Steps in MPE-seq library preparation. Asterisk indicates the presence of aminoallyl-deoxyuridine, incorporated during reverse transcription. (B) Categories of MPE-seq alignments for quantifying the abundance of molecules at each chemical step of splicing for a single splicing event. (C) The correlation between measurements of splice index (SI, calculated as total unspliced reads divided by total spliced reads) for each splicing event is presented for replicate MPE-seq and RNA-seq S. cerevisiae libraries. Libraries were down sampled to five million reads each. The number of splicing events for which at least one premature and mature alignment exists in both replicates is represented by ‘n’. R² values were calculated using linear regression. MPE-seq count data can be found in Table A.2, and RNA-seq count data can be found in Table A.3.

Fig. 3.

Bioanalyzer traces of replicate S. cerevisiae MPE-seq libraries after size selection by: (A) acrylamide gel, or (B) AmpureXP beads. LM and UM denote the lower and upper markers, respectively. PD denotes the peak resulting from primer dimer molecules.