bulk_load_gff.pl(1) Bulk-load a Bio::DB::GFF database from GFF files.


% bulk_load_gff.pl -d testdb dna1.fa dna2.fa features1.gff features2.gff ...


This script loads a Bio::DB::GFF database with the features contained in a list of GFF files and/or FASTA sequence files. You must use the exact variant of GFF described in Bio::DB::GFF. Various command-line options allow you to control which database to load and whether to allow an existing database to be overwritten.

This script differs from bp_load_gff.pl in that it is hard-coded to use MySQL and cannot perform incremental loads. See bp_load_gff.pl for an incremental loader that works with all databases supported by Bio::DB::GFF, and bp_fast_load_gff.pl for a MySQL loader that supports fast incremental loads.


If the filename is given as ``-'' then the input is taken from standard input. Compressed files (.gz, .Z, .bz2) are automatically uncompressed.

FASTA format files are distinguished from GFF files by their filename extensions. Files ending in .fa, .fasta, .fast, .seq, .dna and their uppercase variants are treated as FASTA files. Everything else is treated as a GFF file. If you wish to load -fasta files from STDIN, then use the -f command-line swith with an argument of '-', as in

    gunzip my_data.fa.gz | bp_fast_load_gff.pl -d test -f -

The nature of the bulk load requires that the database be on the local machine and that the indicated user have the ``file'' privilege to load the tables and have enough room in /usr/tmp (or whatever is specified by the \$TMPDIR environment variable), to hold the tables transiently.

Local data may now be uploaded to a remote server via the --local option with the database host specified in the dsn, e.g. dbi:mysql:test:db_host

The adaptor used is dbi::mysqlopt. There is currently no way to change this.

About maxfeature: the default value is 100,000,000 bases. If you have features that are close to or greater that 100Mb in length, then the value of maxfeature should be increased to 1,000,000,000. This value must be a power of 10.

Note that Windows users must use the --create option.

If the list of GFF or fasta files exceeds the kernel limit for the maximum number of command-line arguments, use the --long_list /path/to/files option.


Command-line options can be abbreviated to single-letter options. e.g. -d instead of --database.

   --database <dsn>      Database name (default dbi:mysql:test)
   --adaptor             Adaptor name (default mysql)
   --create              Reinitialize/create data tables without asking
   --user                Username to log in as
   --fasta               File or directory containing fasta files to load
   --long_list           Directory containing a very large number of 
                         GFF and/or FASTA files
   --password            Password to use for authentication
                           (Does not work with Postgres, password must be
                           supplied interactively or be left empty for
                           ident authentication)
   --maxbin              Set the value of the maximum bin size
   --local               Flag to indicate that the data source is local
   --maxfeature          Set the value of the maximum feature size (power of 10)
   --group               A list of one or more tag names (comma or space separated)
                         to be used for grouping in the 9th column.
   --gff3_munge          Activate GFF3 name munging (see Bio::DB::GFF)
   --summary             Generate summary statistics for drawing coverage histograms.
                           This can be run on a previously loaded database or during
                           the load.
   --Temporary           Location of a writable scratch directory


Lincoln Stein, [email protected]

Copyright (c) 2002 Cold Spring Harbor Laboratory

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself. See DISCLAIMER.txt for disclaimers of warranty.