Identifying gene fusions is a cornerstone of modern cancer diagnostics. Gene fusions occur when pieces of two different genes become joined together, often creating abnormal proteins that can drive tumor growth. Some of the most effective targeted cancer therapies exist specifically because these fusion events can be identified and matched to the right drug. However, finding all relevant gene fusions, especially rare ones or those involving unknown gene partners, remains technically challenging.

Many clinical sequencing tests rely on predefined panels that look only for known fusion pairs. While useful, these approaches can miss unexpected fusions or those present at very low levels in a tumor sample. They can also be expensive and inefficient when broader detection is needed. This creates a gap between what researchers know is biologically important and what clinicians can reliably measure.

A team led by researchers at Shanghai Jiao Tong University, has introduced a new sequencing strategy designed to overcome these limitations. Their method, called anchored random reverse primer sequencing, or ARRP-seq, focuses on improving how RNA libraries are prepared before sequencing. Instead of relying solely on fixed primers that target known genes, ARRP-seq uses anchored random reverse primers during PCR. This allows the method to capture gene fusions even when one fusion partner is unknown, while also detecting RNA splicing variants at the same time.

ARRP-seq analyses of gene fusions

a, Schematic overview of ARRP-seq for analysis of unknown gene fusions and comparison with three different panel-based library preparation methods. Most gene fusion analysis methods require fragmentation, ligation and different enrichment techniques downstream. Conversion yields listed here for each step were estimated on the basis of the data reported. Our previously reported multiplex blocker displacement amplification (mBDA) technique can be combined with ARRP-seq to achieve specific enrichment of unknown gene fusions, mutations and RNA splicing variants to improve detection efficiency. b, Simulation analysis of different lengths of RRPs and different inputs of RNA molecules, reflecting the hybridization kinetic information between the RRP and the target template. The heat map scale bar shows the log₁₀ scale of the concentration ratios of RRPs to templates. The red box highlights that the ratio of RRP concentration to the target sample is 1.3-fold to ensure binding of at least one RRP to the template. The heat map colour intensifies towards red as the ratio of RRP to the template increases, indicating a higher rate of reaction. c, To characterize RRP without biased binding on template sequences, sequencing reads were distributed uniformly at the position of 60–150 bases. d, The NGS sequencing coverage distribution of single-plex gene fusion FBXO28-FGFR3 using ARRP-seq analysis. The red coverage pertain to the reads covering exon 3 and exon 4 of FBXO28, and the blue coverage pertain to the reads targeting exon 12 of the FGFR3 gene.

One of the key innovations in this approach is the combination of ARRP-seq with blocker displacement amplification. This additional step selectively enriches rare fusion signals that would otherwise be drowned out by more abundant RNA molecules. As a result, the researchers achieved a median 22-fold enrichment of fusion alleles. In practical terms, this means the method can detect gene fusions at levels about ten times lower than many existing technologies.

Sensitivity is only part of the story. Cost and scalability are equally important for clinical use. The authors report that ARRP-seq reduces sequencing costs by roughly eightfold compared with commonly used fusion detection strategies. This makes it more feasible for routine diagnostic testing and for broader screening efforts where large numbers of samples must be analyzed.

To demonstrate real-world applicability, the researchers applied ARRP-seq to 98 clinical tissue samples. The method identified numerous previously unreported gene fusions, highlighting its ability to uncover novel events that standard panels would miss. In prostate cancer samples, the approach showed strong diagnostic potential by detecting fusion patterns relevant to disease classification. In cervical cancer, it demonstrated promise for personalized diagnostics, where each patient’s tumor may carry a unique combination of genetic changes.

Beyond immediate clinical applications, this work has broader implications for cancer research and drug development. Novel gene fusions can reveal new disease mechanisms and identify potential therapeutic targets. By making fusion detection more sensitive, comprehensive, and affordable, ARRP-seq could accelerate the discovery of clinically actionable alterations across many cancer types.

Overall, this new RNA sequencing-based approach addresses several long-standing challenges in gene fusion detection. By improving sensitivity, lowering costs, and expanding the ability to find unknown fusion partners, ARRP-seq represents an important step toward more precise and personalized cancer diagnostics.

Availability – The code used in this study is available in GitHub at https://github.com/NABMElab/ARRP-seq