NAME
load-go.pl
SYNOPSIS
load-go.pl -d go -h mydbserver -datatype go_ont *.ontologyDESCRIPTION
Loads GO data (ontology files, def files, xref files, assoc files) into a GO database. Will also perform additional housekeeping tasks on database if required
MODULES AND SOFTWARE REQUIRED
You will need the 'xsltproc' executable, which is part of libxslt
(You will have this if you have already installed XML::LibXSLT)
You need to have both go-perl and go-db-perl installed
http://www.godatabase.org/dev contains further details on these two modules
This site also has details on the GO database
ARGUMENTS
-d DBNAME
-h DBSERVER
-datatype FORMAT
(see below)
-schema SCHEMA
by default: godb
Other values: chado
Support for the chado schema is in beta. See http://www.gmod.org/chado
-dbms DRIVER
by default: mysql
other values: Pg
Support for PostgreSQL is in beta
-append
by default this script assumes you are loading a dataset for
the FIRST time. it performs only SQL INSERTs in certain
cases rather than checking with SELECT if it needs to update.
if you are loading the same file for the second time, use
this option. the loading will be slightly slower, but it
will append to existing data
You should use this option if you are loading multiple
ontology files in one go!
-no_optimize
by default, loading will be optimized; certain primary keys in
the db will be cached, and certain tables will be INSERTED straight
into without doing an initial SELECT (the presumption is that these
datatypes would only be loaded once). See L<GO::Handlers::godb> for
details.
If this is turned off, then all data will follow the
SELECT followed by UPDATE or INSERT pattern
This will be slower, but will use less memoty as no cache is required
-no_clear_cache
by default the in-memory cache (which reduced SQL lookups) is cleared
after every single file is loaded. This is to prevent massive caches
when we load all association files in a single command line.
If you have plenty of memory, or aren't loading too many assoc
files you may wish to use this option
-fill_path
(TRUE by default, IF an ontology file is parsed)
populates the graph_path transitive closure table on completion
this option can be used without any files as arguments to
fill the path table in an already term-populated db
-no_fill_path
prevents graph_path table being populated after the ontologies
have been loaded
-fill_count
populates the gene_product_count after all files have been loaded
-add_root
adds an explicit root term
this may be necessary for loading from gene_ontology.obo which
has 3 ontologies - it can be useful to make a fake root term
covering these
NOT FUNCTIONAL - CURRENTLY DONE AUTOMATICALLY
-append
you must use this option if you wish to append to data of the
same type in an already loaded database; it switches off
bulkloading option
-replace
removes all data of the same datatype before loading
-ev
filters based on an evidence type
to filter out IEAs, use the not '!' prefix
-ev '!IEA'
DATATYPES
specify these with the -datatype option
go_ont
A GO ontology file.
After loading is completed, the path/closure table will be built
go_def
A GO.defs definitions file
go_xref
A GO xrefs file; eg ec2go
go_assoc
A gene_associations file
If you also specify the -fill_count option the gene_product_count table will also get populated (this is done at
You can also specify the -ev command to filter out specific evidence codes; for example
load-go.pl -d go -h mydbserver -datatype go-ontology *.ontologyobo
An obo formatted file
HOW IT WORKS
First the input file is converted into its native XML format (eg OBO-XML). That native XML format is transformed to an XML format isomorphic to the GO relational database using an XSLT stylesheet. This transformed XML is then loaded using DBIx::DBStag
SEE ALSO
go-dev/xml/xsl/oboxml_to_godb_prestore.xsl
L<DBIx::DBStag>
L<GO::Parser>NOTES
When loading gene_association files, will split large files into multiple smaller files and load these