A novel bioinformatics pipeline for identification and characterization of fusion transcripts in breast cancer and normal cell lines

Abstract

SnowShoes-FTD, developed for fusion transcript detection in paired-end mRNA-Seq data, employs multiple steps of false positive filtering to nominate fusion transcripts with near 100% confidence. Unique features include: (i) identification of multiple fusion isoforms from two gene partners; (ii) prediction of genomic rearrangements; (iii) identification of exon fusion boundaries; (iv) generation of a 5'-3' fusion spanning sequence for PCR validation; and (v) prediction of the protein sequences, including frame shift and amino acid insertions. We applied SnowShoes-FTD to identify 50 fusion candidates in 22 breast cancer and 9 non-transformed cell lines. Five additional fusion candidates with two isoforms were confirmed. In all, 30 of 55 fusion candidates had in-frame protein products. No fusion transcripts were detected in non-transformed cells. Consideration of the possible functions of a subset of predicted fusion proteins suggests several potentially important functions in transformation, including a possible new mechanism for overexpression of ERBB2 in a HER-positive cell line. The source code of SnowShoes-FTD is provided in two formats: one configured to run on the Sun Grid Engine for parallelization, and the other formatted to run on a single LINUX node. Executables in PERL are available for download from our web site: http://mayoresearch.mayo.edu/mayo/research/biostat/stand-alone-packages.cfm.

Figure 1.

The work flow of the fusion detection algorithm implemented in SnowShoes-FTD.

Figure 2.

PCR validation of candidate fusion products. The PCR primers were designed using the template sequences generated by SnowShoes-FTD. The double-stranded cDNA libraries were constructed using total RNAs from each of the cell lines. The primer sequences and the expected PCR product sizes for each of the fusion candidates were detailed in Supplementary Data S4. (a) The PCR products from 50 fusion candidates with unique isoforms. The fusion candidates were grouped by the cell lines in which the fusion candidates were discovered. (b) The PCR products from five fusion candidates with two fusion isoforms each. Note that there are multiple PCR bands in the lanes for CDK12-TMEM104, and the lowest bands were those from the fusion product.

Figure 3.

The identification of in-frame fusion transcripts and their predicted protein sequences. (a) Staring from the fusion junction spanning reads that aligned to both fusion partner genes, the two junction boundary exons from fusion partner genes A and B are identified; (b) obtaining the IDs and sequences of all exons belonging to the two fusion partner genes A and B based on the curated refFlat file. In this example, Gene A has 7 exons with the third exon as the fusion boundary exon, and gene B has 10 exons with the sixth exon as the fusion boundary exon; (c) obtaining all known transcripts for the two fusion partner genes. Gene A has two known transcripts (A1 and A2) both of which contain the fusion boundary exon. Gene B has 4 known transcripts (B1 → B4) and three of which (B1, B3 and B4) contain the fusion boundary exon. (d) Generating the list of exhaustive fusion transcripts using the known transcripts containing the fusion boundary exons. There are six possible fusion transcripts: A1-B1, A1-B3, A1-B4, A2-B1, A2-B3 and A2-B4. Note that because the differences between the transcripts B1 and B4 are ‘fused out’, the fusion transcript of A1-B1 is identical to that of A1-B4. Similarly, A2-B1 is identical to A2-B4. The fusion transcripts that cause frame shift in gene B are defined as ‘out of frame’, and the ones that did not cause any frame shift will be defined as ‘in frame’ fusions. Each of the in-frame fusions will be translated into amino acid sequences of the fusion proteins.

Figure 4.

Detailed description of ARID1A_MAST2 (a) and WIPF2_ERBB2 (b) fusion transcripts. Using the process described in Figure 3, SnowShoes-FTD uses the RNA sequence of all known transcripts of the fusion partners to predict the sequence of all potential in-frame and out of frame fusion transcripts. Abundance of individual exons for each of the fusion partners, normalized to total exon abundance, was extracted from the mRNA-Seq data.