Note A running version of this web application is available at http://yeastrc.org/compgo_geoduck/pages/goAnalysisForm.jsp These files are provided so that underlying source code doing the analysis may be reviewed.

The Java source files for the web application are available in the web-application directory.

The primary file for performing the analysis is in the org/yeastrc/compgo/geoduck/analysis/AnalyzedGOTreeMaker.java file.

A number of Java libraries required to run the web application, but not directly related to the analysis, are not included in this repository for licensing reasons. These libraries are:

• Several Apache commons libraries: beanutils, collections, dbcp, digester, lang, logging, pool, and validator.
• Java Advanced Imaging libraries.
• MySQL Connector
• yFiles library for generating the graph image (commercial).

The Java code requires a MySQL database with two tables: go_annotation and go_protein_counts. The go_annotation table holds the GO annotations associated with protein accession strings, and go_protein_counts are the cached counts for the number of protenis associated with GO terms. The table definitions are:

CREATE TABLE `go_protein_counts` (
  `go_acc` varchar(255) NOT NULL,
  `count` int(11) DEFAULT NULL,
  PRIMARY KEY (`go_acc`)
) 

CREATE TABLE `go_annotation` (
  `db` varchar(255) NOT NULL,
  `db_object_id` varchar(255) NOT NULL,
  `db_object_symbol` varchar(255) NOT NULL,
  `qualifier` varchar(255) DEFAULT NULL,
  `go_id` varchar(255) NOT NULL,
  `db_reference` varchar(2000) NOT NULL,
  `evidence_code` varchar(255) NOT NULL,
  `with_or_from` varchar(255) DEFAULT NULL,
  `aspect` varchar(255) NOT NULL,
  `db_object_name` varchar(2000) DEFAULT NULL,
  `db_object_synonym` varchar(2000) DEFAULT NULL,
  `db_object_type` varchar(255) NOT NULL,
  `taxon` varchar(255) NOT NULL,
  `date` varchar(255) NOT NULL,
  `assigned_by` varchar(255) NOT NULL,
  KEY `db` (`db`,`db_object_id`),
  KEY `db_object_id` (`db_object_id`),
  KEY `go_id` (`go_id`),
  KEY `qualifier` (`qualifier`),
  KEY `aspect` (`aspect`)
)

The go_annotation table must be pre-populated to perform the analysis. In this study, this table was populated using this method:

1. Download GO annotations from Uniprot (http://www.uniprot.org/downloads). Download gene_association.goa_uniprot
2. Parse file into a temporary go_annotation table The columns in the database table (using the table definition above) and file should match. Can be imported using LOAD DATA INFILE. In this example, data were loaded into compgo_uniprot.go_annotation.
3. The Perl script in perl_scripts/populateGOAnnotations.pl was run, which uses the BLAST mapping of the geoduck proteins to Uniprot to create a go_annotation table containing the geoduck protein accession strings associated with the GO annotations for the Uniprot BLAST hits.