Fusion detection methods

For some panels, Ion Reporter™ Software uses the targeted panel design with software algorithms to detect known and novel gene fusion isoforms. The assay uses the following methods to detect fusions.

Targeted method

In the targeted fusion detection method, panel primers are designed to target specific exon-exon junctions of fusions where the driver gene, the partner gene, and the breakpoint between the driver and the partner gene are known. The sequencing reads are mapped to a reference file that contains only the known gene fusions.

Non‑targeted method

In the non-targeted fusion detection method, the panel primers are used to detect fusions between novel combinations of known diver and partner genes. The sequencing reads are mapped to a broader reference, such as the whole‑exome. Mapping the reads to a broader reference allows for the detection of multiple configurations of driver and partner genes as well as detection of novel breakpoints between the known partner and driver genes.

Exon tiling method

The exon tiling method is a partner agnostic fusion detection method that enables the discovery of novel fusion isoforms and breakpoints. In this method, the primers are designed to target each exon‑exon junction of every driver gene. Each driver gene in the test sample is analyzed individually. After the sequencing reads undergo normalization and baseline correction, the software measures the intragenic 3′ to 5′ expression ratio for each gene and compares the ratio to the baseline (normal sample). Genes that do not undergo a fusion event are expected to have a 3′ to 5′ expression ratio similar to the baseline. Genes that undergo a fusion event typically have a 3′ to 5′ expression ratio greater than the baseline. The imbalance score measures the magnitude of change in 3′ to 5′ expression ratio relative to the baseline. For each driver gene in which fusion was detected, the software also predicts the most likely position of fusion breakpoint. This allows for discovery of novel fusion breakpoints.

Imbalance score = Observed imbalance (test sample) ÷ Expected imbalance (normal sample)

For example, if the observed (test sample) 3′ to 5′ expression ratio is 3, while the expected 3′ to 5′ imbalance for a wild type transcript is 1.5, the imbalance score is 2. Typically, an imbalance score of ≥1.75–2 is indicative of a gene fusion event.

The significance of the expression imbalance is measured by the imbalance p‑value. The p‑value measures the significance of the imbalance at the predicted breakpoint compared to the negative control gene in the sample. Both, the p‑value and the imbalance score are used to determine the occurrence of a fusion event.

Figure 1. Representative primer design for an exon tiling assay
Figure 2. Example coverage profile for a sample with no fusion present

In this example, no fusion is present in the sample. The wild type transcript has uniform coverage of 3′ and 5′ introns.

Figure 3. Example coverage profile for a sample with fusion present

In this example, a mixture of wild type and fusion transcript is present in the sample. The presence of the fusion transcript accounts for the elevated expression of the 3′ gene region.

  1. Predicted fusion breakpoint