Data retrieval

1
Data retrieval
BioMart
Export View
Data sets on ftp site MySQL queries of
databases Perl API access to databases
2
ExportView
3
Data Mining in Ensembl with EnsMart
August 2005
4
Possible queries

• All genes from a candidate region
• Genes with a particular protein domain
• Members of a protein family
• Genes associated with SNPs

5
More specific queries

• Human genes with upstream regions conserved
  w.r.t. mouse
• Upstream sequence for all Ensembl genes mapped to
  U95A chip (similarly, complete genomic annotation
  of MG_U74).
• Genomic location and description of all mouse,
  rat and fugu homologues of all human genes, with
  transmembrane domains, expressed in
  cardiovascular system and have non-synonymous
  SNPs.

6
Ensembl core database

• Normalised
• Each data point stored only once
• Quick updates
• Minimal storage requirements
• But
• Many tables
• Many joins for complicated queries
• Slow for data mining questions

7
BioMart and EnsMart

• Large-scale data retrieval tool
• Query builder interface
• Databases Ensembl, SNP, Vega, (MSD, UniProt)
• Associated features or sequences
• Flexible output formats
• http//www.ebi.ac.uk/biomart/
• http//www.ensembl.org/EnsMart/

8
Mart database

• De-normalised
• Tables with redundant information
• Query-optimised
• Fast and flexible
• designed for data mining

9
Primary Data Sets

• Ensembl genes
• SNP
• Single nucleotide polymorphisms
• Deletion-insertion polymorphisms
• Short tandem repeats
• Vega genes
• (MSD protein structures)
• (UniProt proteomes)

10
Secondary Data Sets

• Markers
• Diseases
• Gene ontology
• Gene expression information
• Homology predictions
• Protein annotation

11
Information flow
start
filter
output
12
BioMart
http//www.biomart.org/
13
BioMart - Features
14
BioMart - Sequences
15
Output formats
16
What about queries not possible to do in EnsMart

• Direct database access at ensembldb.ensembl.org
• martdb.ebi.ac.uk
• MySQL client

Download MySQL for Windows http//www.winmysql.com
/page4.html File wmysr11.zip
17
Access via Perl object API

• Based on bioperl
• Ensembl modules
• For an introduction, see the tutorial at
• http//www.ensembl.org/info/software/core/

18
There are other ways
MartShell
Commandline interface to Mart written in Java.
It works with a Mart Query Language
19
MartExplorer