PSI-Search

Introduction

PSI-Search, similar to PSI-BLAST, combines the Smith-Waterman search algorithm with the PSI-BLAST profile construction strategy to find distantly related protein sequences. Searches are done with SSEARCH, and the selected hits are combined with BLASTPGP to build a position specific scoring matrix (PSSM), which is then used for another search with SSEARCH in the next iteration.

FAQs
Official Website
Download Software

How to use this tool

Running a tool from the web form is a simple multiple steps process, starting at the top of the page and following the steps to the bottom.

Each tool has at least 2 steps, but most of them have more:

  • The first steps are usually where the user sets the tool input (e.g. sequences, databases...)
  • In the following steps, the user has the possibility to change the default tool parameters
  • And finally, the last step is always the tool submission step, where the user can specify a title to be associated with the results and an email address for email notification. Using the submit button will effectively submit the information specified previously in the form to launch the tool on the server

Note that the parameters are validated prior to launching the tool on the server and in the event of a missing or wrong combination of parameters, the user will be notified directly in the form.

Step 1 - Database

Databases

The databases to run the sequence similarity search against. Multiple databases can be selected at the same time.

Database Name Description Abbreviation
UniProt Knowledgebase The UniProt Knowledgebase (UniProtKB) is the central access point for extensive curated protein information, including function, classification, and cross-references. Search UniProtKB to retrieve "everything that is known" about a particular sequence uniprotkb
UniProtKB/Swiss-Prot The manually curated subsection of the UniProt Knowledgebase uniprotkb_swissprot
UniProtKB/Swiss-Prot isoforms The isoform sequences for the manually curated subsection of the UniProt Knowledgebase uniprotkb_swissprotsv
UniProtKB/TrEMBL Subsection of the UniProt Knowledgebase derived from ENA Sequence (formerly EMBL-Bank) coding sequence translations with annotation produced by an automated process. uniprotkb_trembl
UniProtKB Taxonomic Subsets
UniProtKB Archaea Taxonomic subset of the UniProt Knowledgebase for archaea uniprotkb_archaea
UniProtKB Arthropoda Taxonomic subset of the UniProt Knowledgebase for arthropoda uniprotkb_arthropoda
UniProtKB Bacteria Taxonomic subset of the UniProt Knowledgebase for bacteria uniprotkb_bacteria
UniProtKB Complete Microbial Proteomes Taxonomic subset of the UniProt Knowledgebase for complete microbial proteomes uniprotkb_complete_microbial_proteomes
UniProtKB Eukaryota Taxonomic subset of the UniProt Knowledgebase for eukaryota uniprotkb_eukaryota
UniProtKB Fungi Taxonomic subset of the UniProt Knowledgebase for fungi uniprotkb_fungi
UniProtKB Human Taxonomic subset of the UniProt Knowledgebase for human uniprotkb_human
UniProtKB Mammals Taxonomic subset of the UniProt Knowledgebase for mammals uniprotkb_mammals
UniProtKB Nematoda Taxonomic subset of the UniProt Knowledgebase for nematoda uniprotkb_nematoda
UniProtKB PDB Taxonomic subset of the UniProt Knowledgebase for PDB uniprotkb_pdb
UniProtKB Rodents Taxonomic subset of the UniProt Knowledgebase for rodents uniprotkb_rodents
UniProtKB Vertebrates Taxonomic subset of the UniProt Knowledgebase for vertebrates uniprotkb_vertebrates
UniProtKB Viridiplantae Taxonomic subset of the UniProt Knowledgebase for viridiplantae uniprotkb_viridiplantae
UniProtKB Viruses Taxonomic subset of the UniProt Knowledgebase for viruses uniprotkb_viruses
UniProt Clusters The UniProt Reference Clusters (UniRef) databases combine closely related sequences into a single record to speed up searches.
UniProt Clusters 100% The UniProt Reference Clusters (UniRef) containing sequences which are 100% identical. uniref100
UniProt Clusters 100% (SEG filtered) UniProt Reference Clusters database (SEG filtered) with entries that have 100% mutual sequence identity. uniref100_seg
UniProt Clusters 90% The UniProt Reference Clusters (UniRef) containing sequences which are 90% identical. uniref90
UniProt Clusters 50% The UniProt Reference Clusters (UniRef) containing sequences which are 50% identical. uniref50
Patents
EPO Patent Protein Sequences Protein sequences appearing in patents from the European Patent Office (EPO) epop
JPO Patent Protein Sequences Protein sequences appearing in patents from the Japanese Patent Office (JPO) jpop
KIPO Patent Protein Sequences Protein sequences appearing in patents from the Korean Intelectual Property Office (KIP0). kpop
USPTO Patent Protein Sequences Protein sequences appearing in patents from the United States Patent and Trademark Office (USPTO) uspop
NR Patent Proteins Level-1 Non-redundant Patent Protein sequences Level 1 covering data from the EPO, JPO, KIPO and USPTO nrpl1
NR Patent Proteins Level-2 Non-redundant Patent Protein sequences Level 2 covering data from the EPO, JPO, KIPO and USPTO nrpl2
Structure
Protein Structure Sequences Protein sequences from structures described in the Brookhaven Protein Data Bank (PDB) pdb
Structural Genomics Targets Structural Genomic Targets (SGT) database sgt
Others Protein Databases
UniProt Archive The UniProt Archive (UniParc) contains available protein sequences collected from many different sources. The sequence data are archived to facilitate examination of changes to sequence data. Search UniParc if you want to examine the "history" of a particular sequence. uniparc
IntAct The IntAct sequence database is derived from UniProt entries and data from MassSpec experiments submitted to the IntAct protein-interaction database. intact
IMGT/HLA The human major histocompatibility complex (HLA) section of the the international immunogenetics (IMGT) database. imgthlap
IPD-KIR Human Killer-cell Immunoglobulin-like Receptors (KIR) sequence in the Immuno Polymorphism Database (IPD) ipdkirp
IPD-MHC Major Histocompatibility Complex (MHC) section of the Immuno Polymorphism Database (IPD) ipdmhcp

Step 2 - Sequence

Sequence Input Window

The query sequence can be entered directly into this form. The sequence can be be in GCG, FASTA, EMBL, GenBank, PIR, NBRF, PHYLIP or UniProtKB/Swiss-Prot format. A partially formatted sequence is not accepted. Adding a return to the end of the sequence may help certain applications understand the input. Note that directly using data from word processors may yield unpredictable results as hidden/control characters may be present.

Sequence File Upload

A file containing a valid sequence in any format (GCG, FASTA, EMBL, GenBank, PIR, NBRF, PHYLIP or UniProtKB/Swiss-Prot) can be used as input for the sequence similarity search. Word processors files may yield unpredictable results as hidden/control characters may be present in the files. It is best to save files with the Unix format option to avoid hidden Windows characters.

Step 3 - Parameters

Matrix

The comparison matrix to be used to score alignments when searching the database

Matrix Name Abbreviation
BLOSUM45 BLOSUM45
BLOSUM62 BLOSUM62
BLOSUM80 BLOSUM80
PAM30 PAM30
PAM70 PAM70

Default value is: BLOSUM62

Additional information

Gap Open Penalty

Penalty taken away from the score when a gap is created in sequence. Increasing the gap opening penalty will decrease the number of gaps in the final alignment.

Default value is: 11

Additional information

Gap Extend Penalty

Penalty taken away from the score for each base or residue in the gap. Increasing the gap extension penalty favours short gaps in the final alignment, conversly, decreasing the gap extension penalty favours long gaps in the final alignment.

Default value is: 1

Additional information

HOE Region Masking

Turn on/off the sequence masking for HOEs in PSSM constructions. This option allows you to mask sequence characters beyond the alignment region when constructing the PSSM, reducing over-extension errors.

Default value is: yes [true]

Results E() Limit

Limits the number of scores and alignments reported based on the expectation value. This value is the maximum number of times the match is expected to occur by chance.

Default value is: 10.0

PSSM E-Value cut-off

Expectation value threshold for automatic selection of matched sequences for inclusion in the PSSM at each iteration.

Default value is: 1.0e-3

Scores

Maximum number of alignment score summaries reported in the result output.

Default value is: 500

Alignments

Maximum number of alignments reported in the result output.

Default value is: 500

Multi HSPs

Turn on/off the display of all significant alignments between query and database sequence.

Default value is: no [false]

Score Format

Different score formats.

Default value is: Default [default]

Filter

Filter regions of low sequence complexity. This can avoid issues with low complexity sequences where matches are found due to composition rather then meaningful sequence similarity. However in some cases filtering also masks regions of interest and so should be used with caution.

Program Name Description Abbreviation
none No filtering of the query sequence. none
seg Uses the SEQ filter (Wootton and Federhen, 1993) to replace low-complexity regions with 'X' in protein query sequences. seg
xnu Uses the XNU filter (Claverie and States, 1993) to mask statistically significant tandem repeats in protein query sequences. xnu
seg+xnu Uses both SEG and XNU to filter low-complexity regions and statistically significant tandem repeats in protein query sequences. seg+xnu

Default value is: none

Histogram

Turn on/off the histogram in the PSI-Search result. The histogram gives a qualitative view of how well the statistical theory fits the similarity scores calculated by the program.

Default value is: no [false]

Annotation Features

Turn on/off annotation features. Annotation features shows features from UniProtKB, such as variants, active sites, phospho-sites and binding sites that have been found in the aligned region of the database hit. To see the annotation features in the results after this has been enabled, select sequences of interest and click to 'Show' Alignments. This option also enables a new result tab (Domain Diagrams) that highlights domain regions.

Default value is: no [false]

Checkpoint File Upload

Checkpoint file from the previous iteration. Must be in ASN.1 Binary Format.

Step 4 - Submission

Job title

It's possible to identify the tool result by giving it a name. This name will be associated to the results and might appear in some of the graphical representations of the results.

Email Notification

Running a tool is usually an interactive process, the results are delivered directly to the browser when they become available. Depending on the tool and its input parameters, this may take quite a long time. It's possible to be notified by email when the job is finished by simply ticking the box "Be notified by email". An email with a link to the results will be sent to the email address specified in the corresponding text box. Email notifications require valid email addresses.

Email Address

If email notification is requested, then a valid Internet email address in the form joe@example.org must be provided. This is not required when running the tool interactively (The results will be delivered to the browser window when they are ready).

References

PSI-Search: iterative HOE-reduced profile SSEARCH searching.
(2012 Jun 15) Bioinformatics (Oxford, England) 28 (12) :1650-1
Homologous over-extension: a challenge for iterative similarity searches.
(2010 Apr) Nucleic acids research 38 (7) :2177-89
Getting more from less: algorithms for rapid protein identification with multiple short peptide sequences.
(2002 Feb) Molecular & cellular proteomics : MCP 1 (2) :139-47
Rapid and sensitive sequence comparison with FASTP and FASTA.
(1990) Methods in enzymology 183 :63-98
Improved tools for biological sequence comparison.
(1988 Apr) Proceedings of the National Academy of Sciences of the United States of America 85 (8) :2444-8
The EMBL-EBI bioinformatics web and programmatic tools framework.
(2015 Jul 01) Nucleic acids research 43 (W1) :W580-4
A new bioinformatics analysis tools framework at EMBL-EBI.
(2010 Jul) Nucleic acids research 38 (Web Server issue) :W695-9
Analysis Tool Web Services from the EMBL-EBI.
(2013 Jul) Nucleic acids research 41 (Web Server issue) :W597-600

Contact details

Support:

For Support on this service: Please contact EMBL-EBI support at http://www.ebi.ac.uk/support/FASTA

The Author:

William R. Pearson (email: wrp@virginia.edu)
Department of Biochemistry
Box 440, Jordan Hall
U. of Virginia
Charlottesville, VA