Querying biological databases

1
Querying biological databases
2
How to query databases

• Query languages e.g. SQL
• Can query with keyword or phrase
• Boolean queries
• Regular expressions
• Basic database querying is usually done through
  web interface
• Text or sequence-based searches
• Can use Boolean queries and regular expressions

3
Keywords and phrases

• Most searches are case insensitive
• Keywords are single words searched
• Phrases groups of words
• E.g. tyrosine protein kinase returns anything
  with either of the words tyrosine , protein
  or kinase (keywords)
• tyrosine protein kinase returns anything with
  the complete phrase only

4
Boolean operators (George Boole)

• Operators e.g. (AND), (OR), ! (NOT), e.g.
• protein kinase ! tyrosine
• tyrosine protein kinase
• More complex (tyrosine OR kinase) AND (NOT
  serine)
• Operators dont work in , e.g. tyrosine and
  kinase
• Wildcards and ? E.g. cellase finds all words
  starting with cell and ending in ase
• Attributes are used to be more specific about
  where to find the keyword

5
Resources for searching databases

• Each database usually has own web interface
  allowing simple queries
• EBI or NBN Sequence Retrieval System (SRS)
• EB-eye searches all EBI databases
• NCBI Entrez
• E.g. EnsMart allows querying of Ensembl database

6
UniProt website (http//www.uniprot.org/)
7
Sequence Retrieval System

• http//srs.ebi.ac.uk
• Integrates over 150 different databases
• A database can be searched and results linked to
  other databases in SRS
• Searches can be simple or complex
• Results can be viewed in default formats or
  user-defined views
• Users can save their results and launch software
  packages (gt100 applications) to analyse results

8
SRS at EBI -library page
9
Advantages of SRS

• Choice of databases to search
• Linking between databases
• Launching applications from results
• Saving results in different formats
• Simple or very complex queries possible
• Can upload own database

10
NCBI Resources
11
Querying Medline

• Simple query, e.g. for author jones sm
• Using search fields (attributes) jones AU
• More complex query jones AU 1999 DP
• Using Boolean operators gene AND regulation or
  cell AND (repression OR activation)
• If no operators specified, assumes AND

12
Types of search fields

• Title Words TI MeSH Terms MH
• Title/Abstract Words TIAB Language LA
• Text Words TW Journal Title TA
• Substance Name NM Issue IP
• Subset SB Filter FILTER
• Secondary Source ID SI Entrez Date EDAT
• Subheadings SH EC/RN Number RN
• Publication Type PT Author Name AU
• Publication Date DP All Fields ALL
• Personal Name as Subject PS Affiliation AD
• Page Number PG Unique Identifiers UID
• Title Words TI MeSH Major Topic MAJR
• MeSH Date MHDA

13
Tutorial http//cbio.uct.ac.za/training/courses
/short-courses/nbn-databases-course/querying-datab
ases-tutorial-1/