BAR+ web server functionalities are described in Help.
When your sequence is not in BAR+ it is recognised as new by the system and a BLAST alignment over the BAR+ clusters
and singletons starts. When the job is finished you may have the following results:
- Your sequence falls into clusters with Sequence Identity ≥ 40% and Coverage ≥ 90%. If the cluster is endowed
with Pfam and validated GO terms your sequence safely inherits the annotation/s. If the cluster is endowed with
structural template/s your sequence can be modelled with the Cluster-HMM. P0CH37
- Your sequence falls into clusters with Sequence Identity < 40% and Coverage < 90%. If the cluster is endowed with
Pfam and validated GO terms your sequence may still inherit the annotation/s if you manually curate it. If the cluster
is endowed with structural template/s your sequence can be modelled with the Cluster-HMM.
P86346 [BAR+ output]
- Your sequence falls into a singleton with Sequence Identity ≥ 40% and Coverage ≥ 90%. If the singleton carries
Pfam and GO terms your sequence safely inherits the annotation/s and the structure, if available.
D7UPK7 [BAR+ output]
- Your sequence falls into a singleton with Sequence Identity < 40% and Coverage < 90%. No result is returned. Q9YZP2
- BAR+ for functional annotation
Example: a human protein from UniProtKB, accession Q9BW99, with evidence at the transcript level is without any functional annotation, Gene Ontology terms or meaningful keywords. UniProtKB is not linked to UniProtKB-GOA for the prediction of GO terms.
BAR+ solution: submitting the accession code to BAR+, the server returns the page that lists the characteristics of the cluster where the protein is included. It is evident that in this case the protein in endowed with a cluster inherited richer annotation than in the original UniProtKB file [BAR+ output]. In the cluster column are present the validated GO and Pfam terms (P-value < 0.01) and it is clear that the case protein is involved in the ubiquitination related processes, that is localised in the endoplasmic reticulum and that it can bind a carbohydrate. At the bottom of the cluster column there is a link to see the complete list of all data associated with the cluster (see Cluster annotation data format).
- BAR+ for structural annotation
Example: a protein in UniProtKB, accession P30459, is endowed with well annotated functions. It lacks however a PDB structure. A BLAST search against UniProtKB for possible structural templates gives a list of candidates with a maximum coverage of 276 residues (P01891, P01892, P13746, P30443, P05534) when the target sequence length is 365.The best target to template coverage is 75.6%.
BAR+ solution: searching in BAR+ retrieves cluster 331 [BAR+ output] and all UniProtKB templates (P01891, P01892, P13746, P30443, P05534) are found in the same cluster. More to it, a template (2biiA corresponding to P01900) that covers 99.2% of the target sequence length is present in the cluster. The sequence can therefore be modeled by selecting the PIR alignment with template with the highest coverage. The PIR alignment is generated with the Cluster-HMM that models all the PDB structures in the cluster, when this is the case.
Warning: It may occur that the cluster contains only PDB fragments. Example: Q9DGJ8 is endowed with different PDB fragments. In this case BAR+ returns all the sequence target/templates alignments [BAR+ output].