GBrowse Adaptors

From GMOD
(Redirected from GBrowse adaptors)
Jump to: navigation, search

GBrowse has a flexible adaptor (yes, it is spelled that way and is not "adapter") system for running off various types of databases/sources. A common question is "which adaptor should I be using?" This attempts to answer that question.

Adaptor Other required software Roughly how many users Pros Cons
Bio::DB::SeqFeature::Store (use bp_seqfeature_load.pl) MySQL, PostgreSQL, SQLite, BerkeleyDB Many and growing fast. Roughly 4X faster than Bio::DB::GFF for the same data; designed to work with GFF3 Developed for use with GFF3; about 2X slower than Bio::DB::GFF to load a database
Bio::DB::GFF (use bp_load_gff.pl, bp_bulk_load_gff.pl, bp_fast_load_gff.pl) A relational database server: MySQL, PostgreSQL, Oracle, or BerkeleyDB Lots! (Especially MySQL) Quite fast; large user base; Have to use this if your data is in the (now deprecated) GFF2 format. Does not work well with GFF3 formatted data
Bio::DB::Sam (available from CPAN) SAMtools Growing (particularly with GBrowse2) Very fast access to NextGen sequencing data Difficult to use with GBrowse 1.70
Bio::DB::BigWig and Bio::DB::BigWigSet (available from CPAN) UCSC Formats Growing (particularly with GBrowse2) Very fast access to data in bigWig format Difficult to use with GBrowse 1.70
Bio::DB::BigBed (available from CPAN) UCSC Formats Growing (particularly with GBrowse2) Very fast access to data in bigBed format Difficult to use with GBrowse 1.70
Bio::DB::Das::Chado (available from CPAN) PostgreSQL and a Chado schema Relatively few due to the specialized nature of Chado Allows 'live' viewing of the features in a Chado database Slow compared to Bio::DB::GFF
Bio::DB::Das::BioSQL (available from CPAN) MySQL and a BioSQL schema Relatively few due to the small number of BioSQL users Allows 'live' viewing of the features in a BioSQL database Slow compared to Bio::DB::GFF
Memory (ie, flat file database using either Bio::DB::GFF or SeqFeature::Store) None For real servers, none Easy for rapid development and testing Very slow for more than a few thousand features
LuceGene Lucene (searches indexed flat files) Relatively few

Email Threads

There have been some useful email threads on adaptor choices and tradeoffs.