Protein Molecular Function Prediction by Bayesian Phylogenomics

Abstract

We present a statistical graphical model to infer specific molecular function for unannotated protein sequences using homology. Based on phylogenomic principles, SIFTER (Statistical Inference of Function Through Evolutionary Relationships) accurately predicts molecular function for members of a protein family given a reconciled phylogeny and available function annotations, even when the data are sparse or noisy. Our method produced specific and consistent molecular function predictions across 100 Pfam families in comparison to the Gene Ontology annotation database, BLAST, GOtcha, and Orthostrapper. We performed a more detailed exploration of functional predictions on the adenosine-5′-monophosphate/adenosine deaminase family and the lactate/malate dehydrogenase family, in the former case comparing the predictions against a gold standard set of published functional characterizations. Given function annotations for 3% of the proteins in the deaminase family, SIFTER achieves 96% accuracy in predicting molecular function for experimentally characterized proteins as reported in the literature. The accuracy of SIFTER on this dataset is a significant improvement over other currently available methods such as BLAST (75%), GeneQuiz (64%), GOtcha (89%), and Orthostrapper (11%). We also experimentally characterized the adenosine deaminase from Plasmodium falciparum, confirming SIFTER's prediction. The results illustrate the predictive power of exploiting a statistical model of function evolution in phylogenomic problems. A software implementation of SIFTER is available from the authors.

Exploring the structure and function paradigm

“…Pfam) [64;68;72]. Some attempt to bolster confidence in functional assignments through homology by examining the similarity of functions from multiple relatives identified by BLAST [73;74] or apply 'phylogenomics' using Bayesian approaches to combine information in close branches of a phylogenetic tree [75].…”

Recombinant expression of twelve evolutionarily diverse subfamily Iα aminotransferases

“…Alternative methods of substrate specificity assignment have been developed that do not rely on overall sequence similarity (e.g., SIFTER [15]), but aminotransferases have proven to be particularly challenging due to multiple, independent specificity swapping points throughout evolutionary history of this family (Barbara Engelhardt and Michael Jordan, personal communication). Further, there is experimental evidence from both rational redesign and directed evolution for alternate specificity determinants [16][17][18].…”

Annotating proteins with generalized functional linkages

“…Other Bayesian methods have used confidence in GO evidence codes (19), but we are not aware of another that quantitatively estimates them. Our broad assay reflects the expectation that small-scale laboratory experiments provide the best inferences of protein function (19), but further corroboration of the confidence values (Table S2) with other types of functional linkages is needed.…”

“…Other Bayesian methods have used confidence in GO evidence codes (19), but we are not aware of another that quantitatively estimates them. Our broad assay reflects the expectation that small-scale laboratory experiments provide the best inferences of protein function (19), but further corroboration of the confidence values (Table S2) with other types of functional linkages is needed. Annotations supported by the GO code ''inferred by unreviewed electronic annotation'' (IEA) had not been incorporated by the Saccharomyces Genome Database (SGD) (20) at the time of our data collection.…”