SeedServer expects the input to consist of full length protein(s) and requires user to find out the UniProt ID (e.g. B0LUZ4), so please refer to UniProt for searching.
SeedServer can limit the searches to any chosen clade. We strongly recommend users to limit the search to save computing time. Conversely, the output will equally filtered, saving post-processing of the information by the user.
Default parameters for the component software SeedLinkage are recommended, They stand for:
R: number of recruited paralogs that will occupy the Seed position during recruitment. Increase if number of paralogs are larger than 50.
C: required Seed coverage cutoff. Default is 50%.
S: percentage of sequence similarity cutoff. Default is 50%.
L: fraction of score BLASTp self-score. Used to limit paralog retrieval in Seed organism when a BBH with other organisms are not established with the Seed.
E-value cutoff: Default is 1e-10.
SOV threshold: Does not limit recruitment.
Fundaments
SeedServer relies on two components. Initially Seeds are submitted to SeedLinkage software, which will create homologous groups searching a database comprising UniProt full length proteins. Note that SeedLinkage may generate more than a single cluster if more than one Seed is provided. Subsequently, SeedServer consults a local copy of Kegg Orthology (KO) database which has been enriched by UniRef_50 clusters - UEKO: UniRef enriched KO. If SeedLInkage genetates more than one cluster, they will all be enlarged with KO and UEKO. Moreover, initially each Seed is used as both query and subject for determination of PSI-BLAST self-score, followed by PSI-BLAST iterations using Seed as query and all sequences recruited either by SeedLinkage, KO or UEKO as database. Iterations are limited such that the score obtained for the Seed is not less than 50% of its self-score determined before. PSI-BLAST searches are also limited by E-value (default is 1e-10). Furthermore, all proteins are submitted to secondary structure predictions and the best result for structural overlapping with any Seed (SOV parameter) is reported.
Usage
Depending on the number of Seed sequences, clustering might spend from minutes to hours. Administrators can be reached at seedserver@gmail.com for helpdesk with large jobs.
Developed by Rafael Guedes
Questions please send to rafaelmguedes@ufmg.br
Please cite: ...
