CATH FunFHMMer web server: protein functional annotations using functional family assignments.
Bottom Line: Our server also provides valuable information for the prediction of functional sites.The predictive power of FunFHMMer has been validated on a set of 95 proteins where FunFHMMer performs better than BLAST, Pfam and CDD.Recent validation by an independent international competition ranks FunFHMMer as one of the top function prediction methods in predicting GO annotations for both the Biological Process and Molecular Function Ontology.
Affiliation: Institute of Structural and Molecular Biology, UCL, Darwin Building, Gower Street, WC1E 6BT, UK firstname.lastname@example.org.Show MeSH
Mentions: The FunFHMMer web server takes a protein sequence in FASTA format or UniProt/GenBank sequence identifiers as input. The input sequence is scanned against the library of CATH FunFam HMMs using HMMER3 (24). The results are then collapsed into a single set of CATH domain architectures assigned using DomainFinder3 (16) (Figure 1B). Regions on the query sequence are assigned to a FunFam if the E-value of the match to the HMM is significant (i.e. lower than the maximum E-value that is obtained by scanning each sequence within a particular FunFam against that FunFam's HMM (Figure 1C). This maximum E-value is known as the inclusion threshold of a FunFam. The GO annotations for a matching FunFam are then transferred to the query sequence along with confidence scores that are calculated by taking into consideration the frequency of each GO term among the annotated sequences of the particular FunFam. The confidence scores are then propagated up the GO hierarchy or directed acyclic graph (DAG). Finally, a nonredundant set of GO annotations, each GO term retaining its highest confidence score from all the domain regions, making up the GO annotations for the query protein sequence (Figure 1D).
Affiliation: Institute of Structural and Molecular Biology, UCL, Darwin Building, Gower Street, WC1E 6BT, UK email@example.com.