- As biology has increasingly turned into a data-rich science, the need for storing and communicating large datasets has grown tremendously.
- The obvious examples are the nucleotide sequences, the protein sequences, and the 3D structural data produced by X-ray crystallography and macromolecular NMR.
- The biological information of proteins is available as sequences and structures. Sequences are represented in a single dimension whereas the structure contains the three-dimensional data of sequences.
- A biological database is a collection of data that is organized so that its contents can easily be accessed, managed, and updated.
- A protein database is one or more datasets about proteins, which could include a protein’s amino acid sequence, conformation, structure, and features such as active sites.
- Protein databases are compiled by the translation of DNA sequences from different gene databases and include structural information. They are an important resource because proteins mediate most biological functions.
Importance of Protein Databases
Huge amounts of data for protein structures, functions, and particularly sequences are being generated. Searching databases are often the first step in the study of a new protein. It has the following uses:
- Comparison between proteins or between protein families provides information about the relationship between proteins within a genome or across different species and hence offers much more information that can be obtained by studying only an isolated protein.
- Secondary databases derived from experimental databases are also widely available. These databases reorganize and annotate the data or provide predictions.
- The use of multiple databases often helps researchers understand the structure and function of a protein.
Primary databases of Protein
The PRIMARY databases hold the experimentally determined protein sequences inferred from the conceptual translation of the nucleotide sequences. This, of course, is not experimentally derived information, but has arisen as a result of interpretation of the nucleotide sequence information and consequently must be treated as potentially containing misinterpreted information. There is a number of primary protein sequence databases and each requires some specific consideration.
a. Protein Information Resource (PIR) – Protein Sequence Database (PIR-PSD):
- The PIR-PSD is a collaborative endeavor between the PIR, the MIPS (Munich Information Centre for Protein Sequences, Germany) and the JIPID (Japan International Protein Information Database, Japan).
- The PIR-PSD is now a comprehensive, non-redundant, expertly annotated, object-relational DBMS.
- A unique characteristic of the PIR-PSD is its classification of protein sequences based on the superfamily concept.
- The sequence in PIR-PSD is also classified based on homology domain and sequence motifs.
- Homology domains may correspond to evolutionary building blocks, while sequence motifs represent functional sites or conserved regions.
- The classification approach allows a more complete understanding of sequence function-structure relationship.
- The other well known and extensively used protein database is SWISS-PROT. Like the PIR-PSD, this curated proteins sequence database also provides a high level of annotation.
- The data in each entry can be considered separately as core data and annotation.
- The core data consists of the sequences entered in common single letter amino acid code, and the related references and bibliography. The taxonomy of the organism from which the sequence was obtained also forms part of this core information.
- The annotation contains information on the function or functions of the protein, post-translational modification such as phosphorylation, acetylation, etc., functional and structural domains and sites, such as calcium binding regions, ATP-binding sites, zinc fingers, etc., known secondary structural features as for examples alpha helix, beta sheet, etc., the quaternary structure of the protein, similarities to other protein if any, and diseases that may arise due to different authors publishing different sequences for the same protein, or due to mutations in different strains of an described as part of the annotation.
TrEMBL (for Translated EMBL) is a computer-annotated protein sequence database that is released as a supplement to SWISS-PROT. It contains the translation of all coding sequences present in the EMBL Nucleotide database, which have not been fully annotated. Thus it may contain the sequence of proteins that are never expressed and never actually identified in the organisms.
c. Protein Databank (PDB):
- PDB is a primary protein structure database. It is a crystallographic database for the three-dimensional structure of large biological molecules, such as proteins.
- In spite of the name, PDB archive the three-dimensional structures of not only proteins but also all biologically important molecules, such as nucleic acid fragments, RNA molecules, large peptides such as antibiotic gramicidin and complexes of protein and nucleic acids.
- The database holds data derived from mainly three sources: Structure determined by X-ray crystallography, NMR experiments, and molecular modeling.
Secondary Databases of Protein
The secondary databases are so termed because they contain the results of analysis of the sequences held in primary databases. Many secondary protein databases are the result of looking for features that relate different proteins. Some commonly used secondary databases of sequence and structure are as follows:
- A set of databases collects together patterns found in protein sequences rather than the complete sequences. PROSITE is one such pattern database.
- The protein motif and pattern are encoded as “regular expressions”.
- The information corresponding to each entry in PROSITE is of the two forms – the patterns and the related descriptive text.
- In the PRINTS database, the protein sequence patterns are stored as ‘fingerprints’. A fingerprint is a set of motifs or patterns rather than a single one.
- The information contained in the PRINT entry may be divided into three sections. In addition to entry name, accession number and number of motifs, the first section contains cross-links to other databases that have more information about the characterized family.
- The second section provides a table showing how many of the motifs that make up the fingerprint occurs in the how many of the sequences in that family.
- The last section of the entry contains the actual fingerprints that are stored as multiple aligned sets of sequences, the alignment is made without gaps. There is, therefore, one set of aligned sequences for each motif.
- MHCPep is a database comprising over 13000 peptide sequences known to bind the Major Histocompatibility Complex of the immune system.
- Each entry in the database contains not only the peptide sequence, which may be 8 to 10 amino acid long but in addition has information on the specific MHC molecules to which it binds, the experimental method used to assay the peptide, the degree of activity and the binding affinity observed , the source protein that, when broken down gave rise to this peptide along with other, the positions along the peptide where it anchors on the MHC molecules and references and cross-links to other information.
- Pfam contains the profiles used using Hidden Markov models.
- HMMs build the model of the pattern as a series of the match, substitute, insert or delete states, with scores assigned for alignment to go from one state to another.
- Each family or pattern defined in the Pfam consists of the four elements. The first is the annotation, which has the information on the source to make the entry, the method used and some numbers that serve as figures of merit.
- The second is the seed alignment that is used to bootstrap the rest of the sequences into the multiple alignments and then the family.
- The third is the HMM profile.
- The fourth element is the complete alignment of all the sequences identified in that family.
- Xiong J. (2006). Essential Bioinformatics. Texas A & M University. Cambridge University Press.
- Arthur M Lesk (2014). Introduction to bioinformatics. Oxford University Press. Oxford, United Kingdom