NAME
Bio::SeqIO::chadoxml - chadoxml sequence output stream
SYNOPSIS
It is probably best not to use this object directly, but rather go through the SeqIO handler system:
$writer = Bio::SeqIO->new(-file => ">chado.xml",
-format => 'chadoxml');
# assume you already have a Sequence object
$writer->write_seq($seq_obj);DESCRIPTION
This object can transform Bio::Seq objects to chadoxml flat file databases (for chadoxml DTD, see http://gmod.cvs.sourceforge.net/gmod/schema/chado/dat/chado.dtd).
This is currently a write-only module.
$seqio = Bio::SeqIO->new(-file => '>outfile.xml',
-format => 'chadoxml');
# we have a Bio::Seq object $seq which is a gene located on
# chromosome arm 'X', to be written out to chadoxml
# before converting to chadoxml, $seq object B<must> be transformed
# so that all the coordinates in $seq are against the source
# feature to be passed into Bio::SeqIO::chadoxml->write_seq()
# -- chromosome arm X in the example below.
$seqio->write_seq(-seq=>$seq,
-seq_so_type=>'gene',
-src_feature=>'X',
-src_feat_type=>'chromosome_arm',
-nounflatten=>1,
-is_analysis=>'true',
-data_source=>'GenBank');The chadoxml output of Bio::SeqIO::chadoxml->write_seq() method can be passed to the loader utility in XORT package (http://gmod.cvs.sourceforge.net/gmod/schema/XMLTools/XORT/) to be loaded into chado.
This object is currently implemented to work with sequence and annotation data from whole genome projects deposited in GenBank. It may not be able to handle all different types of data from all different sources.
In converting a Bio::Seq object into chadoxml, a top-level feature is created to represent the object and all sequence features inside the Bio::Seq object are treated as subfeatures of the top-level feature. The Bio::SeqIO::chadoxml object calls Bio::SeqFeature::Tools::Unflattener to unflatten the flat feature list contained in the subject Bio::Seq object, to build gene model containment hierarchy conforming to chado central dogma model: gene --> mRNA --> exons and protein.
Destination of data in the subject Bio::Seq object $seq is as following:
*$seq->display_id: name of the top-level feature;
*$seq->accession_number: if defined, uniquename and
feature_dbxref of the top-level
feature if not defined,
$seq->display_id is used as the
uniquename of the top-level feature;
*$seq->molecule: transformed to SO type, used as the feature
type of the top-level feature if -seq_so_type
argument is supplied, use the supplied SO type
as the feature type of the top-level feature;
*$seq->species: organism of the top-level feature;
*$seq->seq: residues of the top-level feature;
*$seq->is_circular, $seq->division: feature_cvterm;
*$seq->keywords, $seq->desc, comments: featureprop;
*references: pub and feature_pub;
medline/pubmed ids: pub_dbxref;
comments: pubprop;
*feature "source" span: featureloc for top-level feature;
*feature "source" db_xref: feature_dbxref for top-level feature;
*feature "source" other tags: featureprop for top-level feature;
*subfeature 'symbol' or 'label' tag: feature uniquename, if
none of these is present, the chadoxml object
generates feature uniquenames as:
<gene>-<feature_type>-<span>
(e.g. foo-mRNA--1000..3000);
*gene model: feature_relationship built based on the
containment hierarchy;
*feature span: featureloc;
*feature accession numbers: feature_dbxref;
*feature tags (except db_xref, symbol and gene): featureprop;Things to watch out for:
*chado schema change: this version works with the chado
version tagged chado_1_01 in GMOD CVS.
*feature uniquenames: especially important if using XORT
loader to do incremental load into
chado. may need pre-processing of the
source data to put the correct
uniquenames in place.
*pub uniquenames: chadoxml->write_seq() has the FlyBase policy
on pub uniquenames hard-coded, it assigns
pub uniquenames in the following way: for
journals and books, use ISBN number; for
published papers, use MEDLINE ID; for
everything else, use FlyBase unique
identifier FBrf#. need to modify the code to
implement your policy. look for the comments
in the code.
*for pubs possibly existing in chado but with no knowledge of
its uniquename:put "op" as "match", then need to run the
output chadoxml through a special filter that
talks to chado database and tries to find the
pub by matching with the provided information
instead of looking up by the unique key. after
matching, the filter also resets the "match"
operation to either "force" (default), or
"lookup", or "insert", or "update". the
"match" operation is for a special FlyBase use
case. please modify to work according to your
rules.
*chado initialization for loading:
cv & cvterm: in the output chadoxml, all cv's and
cvterm's are lookup only. Therefore,
before using XORT loader to load the
output into chado, chado must be
pre-loaded with all necessary CVs and
CVterms, including "SO" , "property
type", "relationship type", "pub type",
"pubprop type", "pub relationship type",
"sequence topology", "GenBank feature
qualifier", "GenBank division". A pub by
the uniquename 'nullpub' of type 'null
pub' needs to be inserted.FEEDBACK
Mailing Lists
User feedback is an integral part of the evolution of this and other Bioperl modules. Send your comments and suggestions preferably to one of the Bioperl mailing lists. Your participation is much appreciated.
bioperl-l@bioperl.org - General discussion
http://bioperl.org/wiki/Mailing_lists - About the mailing listsReporting Bugs
Report bugs to the Bioperl bug tracking system to help us keep track the bugs and their resolution. Bug reports can be submitted via the web:
http://bugzilla.bioperl.orgAUTHOR - Peili Zhang
Email peili@morgan.harvard.edu
APPENDIX
The rest of the documentation details each of the object methods. Internal methods are usually preceded with a _
write_seq
Title : write_seq
Usage : $stream->write_seq(-seq=>$seq, -seq_so_type=>$seqSOtype,
-src_feature=>$srcfeature,
-src_feat_type=>$srcfeattype,
-nounflatten=>0 or 1,
-is_analysis=>'true' or 'false',
-data_source=>$datasource)
Function: writes the $seq object (must be seq) into chadoxml.
Current implementation:
1. for non-mRNA records,
a top-level feature of type $seq->alphabet is
generated for the whole GenBank record, features listed
are unflattened for DNA records to build gene model
feature graph, and for the other types of records all
features in $seq are treated as subfeatures of the top-level
feature.
2. for mRNA records,
if a 'gene' feature is present, it B<must> have a /symbol
or /label tag to contain the uniquename of the gene. a top-
level feature of type 'gene' is generated. the mRNA is written
as a subfeature of the top-level gene feature, and the other
sequence features listed in $seq are treated as subfeatures
of the mRNA feature.
Returns : 1 for success and 0 for error
Args : A Bio::Seq object $seq, optional $seqSOtype, $srcfeature,
$srcfeattype, $nounflatten, $is_analysis and $data_source.
when $srcfeature (a string, the uniquename of the source
feature) is given, the location and strand information of
the top-level feature against the source feature will be
derived from the sequence feature called 'source' of the
$seq object, a featureloc record is generated for the top
-level feature on $srcfeature. when $srcfeature is given,
$srcfeattype must also be present. All feature coordinates
in $seq should be against $srcfeature. $seqSOtype is the
optional SO term to use as the type of the top-level feature.
For example, a GenBank data file for a Drosophila melanogaster
genome scaffold has the molecule type of "DNA", when
converting to chadoxml, a $seqSOtype argument of
"golden_path_region" can be supplied to save the scaffold
as a feature of type "golden_path_region" in chadoxml, instead
of "DNA". a feature with primary tag of 'source' must be
present in the sequence feature list of $seq, to decribe the
whole sequence record.