IsoSVM – Distinguishing isoforms and paralogs on the protein level
2006
IsoSVM – A Tool for Identifying Protein Isoforms
Sample size: 3802
publication
Evidence: high
Author Information
Author(s): Michael Spitzer, Stefan Lorkowski, Paul Cullen, Alexander Sczyrba, Georg Fuellen
Primary Institution: University of Münster
Hypothesis
Can a support vector machine classifier accurately distinguish between protein isoforms and paralogs without genomic data?
Conclusion
The IsoSVM tool can accurately distinguish isoforms from paralogs with over 97% accuracy.
Supporting Evidence
- The SVM classifier achieved a mean accuracy of 99.55%.
- IsoSVM classified 986 out of 5,459 sequence pairs as isoforms.
- The classifier showed a precision of 99.23% in a manual review of samples.
Takeaway
The IsoSVM tool helps scientists tell the difference between similar proteins, which is important for understanding how genes work.
Methodology
The study used a support vector machine classifier trained on features derived from protein sequence alignments.
Limitations
The classifier may misclassify some paralogs as isoforms due to high similarity.
Statistical Information
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website