Background

The well-documented transcriptome complexity can be viewed as a combination of several hallmark features, including the multitude of alternative splicing events [1,2], multiple alternative transcription start and termination sites (TSSs and TTSs) [3,4], overlapping or chimeric transcripts [5–9], and pervasive transcription [10,11]. Collectively, the molecular processes that give rise to these features orchestrate the sophisticated landscape of mammalian gene expression.

Comprehensive characterization of complex transcriptome landscapes presents several challenges, particularly in detecting low-abundance transcripts, accurately defining transcript structures, and efficiently obtaining full-length sequences. While conventional transcriptome studies effectively identify and annotate highly expressed, ubiquitously present transcripts, they frequently miss elusive transcripts exhibiting tightly regulated expression patterns, produced only in specific tissues, cell types, or under specific biological conditions [12–16]. The restricted expression patterns of these transcripts result in extremely low abundance that often falls below the sensitivity limits of standard RNA-seq assays, hindering their effective detection.

Given these limitations, targeted RNA enrichment approaches have become essential for accessing low-abundance transcripts, especially in bulk samples. Currently, two primary targeted RNA enrichment strategies are employed: CaptureSeq and rapid amplification of cDNA ends (RACE). CaptureSeq, a hybridization-based method, utilizes specifically designed DNA oligonucleotide probes to capture transcripts of interest, thereby increasing sequencing coverage for target regions [7]. This method enhances the sensitivity and accuracy of low-abundance transcript detection and has been applied in the discovery of novel genes and transcripts [7,17]. Moreover, combining CaptureSeq with long-read sequencing platforms like Nanopore [18] or PacBio [19] enables high-precision sequencing without the need for transcript assembly. However, its implementation is costly and highly dependent on probe design accuracy.

RACE, an inverse PCR-based technique, amplifies the 5' and 3' terminal sequences of RNA molecules by targeting specific anchor sequences [20,21]. 3' RACE utilizes a poly(dT) primer to target the poly(A) tail, facilitating the retrieval of 3' terminal sequences [22]. 5' RACE employs various strategies to amplify the 5' end, including terminal transferase tailing, adapter ligation, and the switching mechanism at the 5' end of the RNA template (SMART) approach [22,23]. RACE is a cost-effective and readily accessible technique for most molecular biology laboratories. In a direct comparison study, this method demonstrated higher sensitivity than CaptureSeq in detecting splice junctions [24]. RACE has been combined with tiling arrays or next-generation sequencing (NGS) to resolve complex transcriptional patterns. Kapranov et al. integrated RACE with high-density tiling arrays, revealing the extensive complexity of the human transcriptome, including transcript fusion and interlacing structures [5]. Lagarde et al. developed RACE-Seq, performing 5' and 3' RACE on 398 known long noncoding RNA (lncRNA) exons, followed by high-throughput sequencing using the Roche 454 FLX+ NGS platform, yielding reads with an average length of approximately 600 base pairs (bp) [24]. In our recent studies, RACE was integrated with Nanopore long-read sequencing to determine the complete structure of novel intragenic or intergenic transcripts, using GENSCAN-predicted exons as anchors [25,26]

The inherent complexity of the transcriptome, characterized by dynamic splicing patterns and structural diversity, poses significant analytical challenges. Conventional RACE coupled with Sanger sequencing, while capable of generating relatively long reads, suffers from low throughput and labor-intensive workflows. Conversely, short-read NGS technologies have intrinsic limitations in adequately resolving complex splicing patterns. Although long-read sequencing overcomes read length limitations and enables full-length transcript detection [27], its modest throughput remains a bottleneck for capturing the full diversity of low-abundance transcripts. Therefore, combining targeted RNA enrichment techniques with long-read sequencing offers an effective strategy, providing a practical and streamlined solution for targeted analysis of complex loci. Among long-read sequencing techniques, Nanopore sequencing achieves significantly longer read lengths than PacBio and is more cost-effective, providing unprecedented capability for de novo gene annotation and structural variant detection [19,28,29].

Here, we introduce RACE coupled with Nanopore sequencing (RACE-Nano-Seq), a method designed to efficiently capture full-length transcripts. This approach leverages target locus sequences (annotated exons, predicted exons, or other sequences) as anchors for full-length transcript enrichment via 5'/3' RACE. The enriched cDNA products are then analyzed with Nanopore sequencing and aligned to the corresponding reference genome to enable sensitive transcriptome characterization. This approach is particularly well-suited for detecting low-abundance transcripts, identifying novel exons, and characterizing splicing patterns at specific gene loci. In this protocol, we detail the experimental procedures and analytical pipelines for implementing RACE-Nano-Seq. Additionally, we provide an example demonstrating its application in profiling the transcriptome diversity at a specific gene locus.