deFuse: An Algorithm for Gene Fusion Discovery in Tumor RNA-Seq Data
Author Information
Author(s): Andrew McPherson, Fereydoun Hormozdiari, Abdalnasser Zayed, Ryan Giuliany, Hav G. Sun, Mark G. F. Sun, Malachi Griffith, Alireza Heravi-Moussavi, Janine Senz, Nataliya Melnyk, Marina Pacheco, Marco A. Marra, Martin Hirst, Torsten O. Nielsen, S. Cenk Sahinalp, David Huntsman, Sohrab P. Shah
Primary Institution: BC Cancer Agency
Hypothesis
Can a novel computational method improve the discovery of gene fusions in tumor RNA-Seq data?
Conclusion
The deFuse algorithm successfully identifies gene fusions in ovarian cancer and sarcoma samples with better sensitivity and specificity than existing methods.
Supporting Evidence
- deFuse identified 45 novel gene fusions in ovarian cancer, marking the first reported fusions in this tumor type.
- The classifier trained on 121 examples achieved an AUC of 0.91, indicating high accuracy.
- Validation of 46 predictions showed a high success rate, with 42 confirmed by RT-PCR.
- deFuse outperformed existing methods like FusionSeq and MapSplice in sensitivity and specificity.
- Analysis of gene expression patterns indicated that many fusions could have functional consequences.
Takeaway
Researchers created a new tool called deFuse to find gene fusions in cancer samples, which helps understand how cancers develop.
Methodology
The study used RNA-Seq data from ovarian and sarcoma samples, applying the deFuse algorithm to identify gene fusions.
Potential Biases
The selection of candidates for validation may introduce bias due to the imbalance of positive and negative examples.
Limitations
The method requires at least five discordant read pairs to nominate a gene fusion, which may miss low-expression fusions.
Participant Demographics
The study involved 40 ovarian cancer patients and 3 sarcoma patients.
Statistical Information
P-Value
0.91
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website