Using Linear Algebra for Protein Structural Comparison and Classification
Author Information
Author(s): Gomide Janaína, Melo-Minardi Raquel, dos Santos Marcos Augusto, Neshich Goran, Meira Wagner Jr., Lopes Júlio César, Santoro Marcelo
Primary Institution: Universidade Federal de Minas Gerais
Hypothesis
Can linear algebra be used to effectively compare and classify protein structures into fold families?
Conclusion
The study demonstrates that a novel methodology using linear algebra can classify proteins with a precision of up to 80%.
Supporting Evidence
- The methodology achieved a precision of up to 80% in classifying proteins.
- Contact maps were converted into vectors for analysis.
- Latent Semantic Indexing was used to index the protein database.
Takeaway
This study shows a new way to group proteins based on their shapes using math, helping scientists find similar proteins more easily.
Methodology
The study used linear algebra techniques, specifically Singular Value Decomposition (SVD) and Latent Semantic Indexing (LSI), to analyze protein contact maps.
Limitations
The methodology was only tested with hydrophobic interactions and may not generalize to all types of protein interactions.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website