Method

The SeqFEATURE library of 3D functional site models: comparison to existing methods and applications to protein function annotation

Shirley Wu¹ , Mike P Liang² and Russ B Altman^1,3,4

¹Program in Biomedical Informatics, Stanford University, Stanford, CA, 94305 USA

²Google, Inc., Amphitheatre Pkwy, Mountain View, CA, 94043 USA

³Department of Genetics, Stanford University, Stanford, CA, 94305 USA

⁴Department of Bioengineering, Stanford University, Stanford, CA, 94305 USA

author email corresponding author email

Genome Biology 2008, 9:R8doi:10.1186/gb-2008-9-1-r8

Published:	16 January 2008

Subject areas: Biochemistry and structural biology, Bioinformatics

Abstract

Structural genomics efforts have led to increasing numbers of novel, uncharacterized protein structures with low sequence identity to known proteins, resulting in a growing need for structure-based function recognition tools. Our method, SeqFEATURE, robustly models protein functions described by sequence motifs using a structural representation. We built a library of models that shows good performance compared to other methods. In particular, SeqFEATURE demonstrates significant improvement over other methods when sequence and structural similarity are low.