Comprehensive annotation of 3′UTRs from primary cells and their quantification from scRNA-seq data

Abstract

Although half of human genes use alternative polyadenylation (APA) to generate mRNA isoforms that encode the same protein but differ in their 3′UTRs, most single cell RNA-sequencing (scRNA-seq) pipelines only measure gene expression. Here, we describe an open-access pipeline, called scUTRquant (https://github.com/Mayrlab/scUTRquant), that measures gene and 3′UTR isoform expression from scRNA-seq data obtained from known cell types in any species. scUTRquant-derived gene and 3′UTR transcript counts were validated against standard methods which demonstrated their accuracy. 3′UTR isoform quantification was substantially more reproducible than previous methods. scUTRquant provides an atlas of high-confidence 3′ end cleavage sites at single-nucleotide resolution to allow APA comparison across mouse datasets. Analysis of 120 mouse cell types revealed that during differentiation genes either change their expression or they change their 3′UTR isoform usage. Therefore, we identified thousands of genes with 3′UTR isoform changes that have previously not been implicated in specific biological processes.

A Survey on Methods for Predicting Polyadenylation Sites from DNA Sequences, Bulk RNA-seq, and Single-cell RNA-seq

“…Initial attempts have been made for APA analysis using multi-omics data. scUTRquant [108] incorporates a cleavage site atlas established from a mouse full-length Microwell-seq dataset of 400,000 single cells [109] for filtering high-confidence pAs predicted from 3′ tag scRNA-seq. Leung, et al [73] predicted strength of pAs using nucleotide sequences, considering features from additional layers like nucleosome positioning and RNA/binding protein motifs.…”

“…In contrast to the peak calling-based methods used for de novo pA identification, a few approaches identify pAs base on prior pA annotations, including MAPPER [106], SCAPTURE [107], and scUTRquant [108]. Li et al developed MAPPER [106] for predicting pAs from both bulk RNA-seq and scRNA-seq data, which incorporates annotated pAs in PolyA_DB 3 [85] and pools single cells of the same type to mimic pseudo-bulk samples.…”

“…The authors used SCAPTURE to profile APA dynamics between COVID-19 patients and healthy individuals, and found the preference of proximal pA usage in numerous immune response-associated genes upon SARS-CoV-2 infection. Fansler et al developed scUTRquant [108] for measuring 3’ UTR isoform expression from scRNA-seq, which relies on a cleavage site atlas established from GENCODE annotation and a mouse Microwell-seq dataset of 400,000 single cells [109].…”

CPA-Perturb-seq: Multiplexed single-cell characterization of alternative polyadenylation regulators

“…In particular, the majority of scRNA-seq protocols are explicitly designed to capture the 3' end of polyadenylated mRNA transcripts. Therefore, these methods are well-suited to quantify transcriptome-wide polyA site usage at single-cell resolution alongside gene abundances, revealing dynamic changes in polyadenylation during cellular differentiation and disease [35][36][37][38] .…”