[med-svn] r2485 - trunk/community/infrastructure/getData

[plessy at alioth.debian.org]

Author: plessy
Date: 2008-09-12 07:00:18 +0000 (Fri, 12 Sep 2008)
New Revision: 2485

Modified:
   trunk/community/infrastructure/getData/getData
Log:
Added Human RefSeq and ENSEMBL.


Modified: trunk/community/infrastructure/getData/getData
===================================================================
--- trunk/community/infrastructure/getData/getData	2008-09-11 15:27:41 UTC (rev 2484)
+++ trunk/community/infrastructure/getData/getData	2008-09-12 07:00:18 UTC (rev 2485)
@@ -3,7 +3,7 @@
 use strict;
 use vars qw(%toBeMirrored $sharedWgetOptions);
 
-$sharedWgetOptions = "--recursive --no-host-directories --no-directories --level 1 --no-parent";
+$sharedWgetOptions = "--continue --recursive --no-host-directories --no-directories --level 1 --no-parent";
 
 =head1 NAME
 
@@ -336,6 +336,11 @@
 #                    B  I  O  I  N  F  O  R  M  A  T  I  C  S
 #
 
+	"ensembl.hsa" => {
+		name => "ENSEMBL flatfile dump -- Homo Sapiens",
+		source => "wget $sharedWgetOptions ftp://ftp.ensembl.org/pub/current_embl/homo_sapiens/*"
+	},
+
 	"gene.ontology.rdf" => {
 		name => "GeneOntology - RDF-formatted, terms only.",
 		source => "wget $sharedWgetOptions ftp://ftp.geneontology.org/pub/go/godatabase/archive/latest/go_*-termdb.rdf-xml.gz",
@@ -395,6 +400,12 @@
 		source => "wget $sharedWgetOptions http://www.reactome.org/download/interactions.README.txt http://www.reactome.org/download/current/homo_sapiens.interactions.txt.gz"
 	},
 
+# Proof-of-principle for RefSeq. Does not include everything.
+	"refseq.hsa" => {
+		name => "The NCBI Reference Sequence project - Homo sapiens",
+		source => "wget $sharedWgetOptions ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/mRNA_Prot/*ff.gz"
+	},
+
 	"pfam-a" => {
 		name => "Pfam-A : Manually curated protein families and domains, only the seed is presented.",
 		source => "wget $sharedWgetOptions ftp://ftp.sanger.ac.uk/pub/databases/Pfam/current_release/Pfam-A-seed.gz"