Prediction of functional specificity determinants from protein sequences using log-likelihood ratios

Abstract

SPEL is freely available for non-commercial use. Its pre-compiled versions for several platforms and alignments used in this work are available at ftp://iole.swmed.edu/pub/SPEL/

Enhanced function annotations for Drosophila serine proteases: A case study for systematic annotation of multi-member gene families

“…A number of computational methods have been developed to identify functionally important sites within protein families. These include approaches that employ parameters such as sequence and phylogenetic patterns (Lichtarge et al, 1996;Pei et al, 2006), 3D structures (Jones and Thornton, 2004;Jambon et al, 2005;Brylinski et al, 2007) and residue physical properties (Elcock, 2001). But none of these approaches explicitly utilize structural classification and function annotation present in sequence databases.…”

“…Their method was able to identify only 4 residues, all present in the C-terminal half of the proteases. Methods like TreeDet (Carro et al, 2006) and SPEL (Pei et al, 2006) could not handle multiple sequence alignment of serine proteases used by us.…”

WRKY gene family evolution in Arabidopsis thaliana

“…The Arabidopsis genome was also scanned using BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) to check whether any sequence was missed due to incomplete or erroneous annotation. BLASTCLUST is much more robust (Pei et al 2006) and therefore, used to cluster the amino acid sequences (Altschul et al 1997). The option of sequence length to be covered was set to 90% and percent identity threshold was set to 30%.…”

Context-Specific Independence Mixture Modelling for Protein Families

“…This is not unexpected as one would expect differing amounts of synonymous substitutions within the various subgroups and that is the situation where the DMP makes the largest difference as compared to the uninformative prior. For comparison we also applied the tree-based method SPEL [22] to our data. The sources were obtained from the authors and run with default parameters.…”

“…In many cases the subgroup structure of a given family is a direct consequence of evolutionary divergence of homologue sequences. As such it is not surprising that methods based on the phylogenetic tree of a family have been extensively and successfully used to study protein family subgroups [15,16,22,25]. However, the performance of these methods does degrade in cases where the evolutionary divergence between subgroups is large.…”