refseq location changed. Plus the rna is now obtained and parsed in genbank format

misc-scripts/xref_mapping/sql/populate_metadata.sql

@@ -62,10 +62,10 @@ INSERT INTO source VALUES (1500, 'MIM2', 1, 'Y', 3);
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1, 9606,'ftp://ftp.ebi.ac.uk/pub/databases/SPproteomes/swissprot_files/proteomes/9606.SPC', '', now(), now(), "UniProtParser");
 
 ##       refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 9606,'ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/mRNA_Prot/human.protein.gpff.gz', '', now(), now(), "RefSeqGPFFParser");
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 9606,'ftp://ftp.ncbi.nih.gov/genomes/H_sapiens/protein/protein.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
 ##       refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 9606,'ftp://ftp.ncbi.nih.gov/refseq/H_sapiens/mRNA_Prot/human.rna.fna.gz', '', now(), now(), "RefSeqParser");
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 9606,'ftp://ftp.ncbi.nih.gov/genomes/H_sapiens/RNA/rna.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
 ##       GO
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1070, 9606,'ftp://ftp.ebi.ac.uk/pub/databases/GO/goa/HUMAN/gene_association.goa_human.gz', '', now(), now(), "GOParser");
@@ -76,18 +76,18 @@ INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date
 ##      Interpro
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1300, 9606,'ftp://ftp.ebi.ac.uk/pub/databases/interpro/short_name.dat ftp://ftp.ebi.ac.uk/pub/databases/interpro/protein2interpro.dat.gz', '', now(), now(), "InterproParser");
 
-
+##      OMIM 
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1500, 9606,'ftp://ftp.ncbi.nih.gov/repository/OMIM/morbidmap', '', now(), now(), "MIMParser");
 
 ###MOUSE
 ##      uniprot
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1, 10090, 'ftp://ftp.ebi.ac.uk/pub/databases/SPproteomes/swissprot_files/proteomes/10090.SPC', '', now(), now(), "UniProtParser");
 
-##      refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10090,'ftp://ftp.ncbi.nih.gov/refseq/M_musculus/mRNA_Prot/mouse.protein.gpff.gz', '', now(), now(), "RefSeqGPFFParser");
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10090,'ftp://ftp.ncbi.nih.gov/genomes/M_musculus/protein/protein.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
-##      refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10090,'ftp://ftp.ncbi.nih.gov/refseq/M_musculus/mRNA_Prot/mouse.rna.fna.gz', '', now(), now(), "RefSeqParser");
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10090,'ftp://ftp.ncbi.nih.gov/genomes/M_musculus/RNA/rna.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
 ##      mgd (MGI -- MarkerSymbol)
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1080, 10090,'ftp://ftp.informatics.jax.org/pub/reports/MRK_SwissProt_TrEMBL.rpt', '', now(), now(), "MGDParser");
@@ -100,11 +100,11 @@ INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date
 ##      uniprot
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1, 10116, 'ftp://ftp.ebi.ac.uk/pub/databases/SPproteomes/swissprot_files/proteomes/10116.SPC', '', now(), now(), "UniProtParser");
 
-##      refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10116,'ftp://ftp.ncbi.nih.gov/refseq/R_norvegicus/mRNA_Prot/rat.protein.gpff.gz', '', now(), now(), "RefSeqGPFFParser");
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10116,'ftp://ftp.ncbi.nih.gov/genomes/R_norvegicus/protein/protein.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
-##      refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10116,'ftp://ftp.ncbi.nih.gov/refseq/R_norvegicus/mRNA_Prot/rat.rna.fna.gz', '', now(), now(), "RefSeqParser");
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 10116,'ftp://ftp.ncbi.nih.gov/genomes/R_norvegicus/RNA/rna.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
 ##      GO 
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1070, 10116,'ftp://ftp.ebi.ac.uk/pub/databases/GO/goa/RAT/gene_association.goa_rat.gz', '', now(), now(), "GOParser");
@@ -117,11 +117,11 @@ INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date
 ##      uniprot
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1, 7955, 'ftp://ftp.ebi.ac.uk/pub/databases/SPproteomes/swissprot_files/proteomes/7955.SPC', '', now(), now(), "UniProtParser");
 
-##      refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 7955,'ftp://ftp.ncbi.nih.gov/refseq/D_rerio/mRNA_Prot/zebrafish.protein.gpff.gz', '', now(), now(), "RefSeqGPFFParser");
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 7955,'ftp://ftp.ncbi.nih.gov/genomes/D_rerio/protein/protein.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
-##      refseq
-INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 7955,'ftp://ftp.ncbi.nih.gov/refseq/D_rerio/mRNA_Prot/zebrafish.rna.fna.gz', '', now(), now(), "RGDParser");
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 7955,'ftp://ftp.ncbi.nih.gov/genomes/D_rerio/RNA/rna.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
 ##      GO 
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1070, 7955,'ftp://ftp.geneontology.org/pub/go/gene-associations/gene_association.zfin.gz', '', now(), now(), "GOParser");
@@ -134,8 +134,11 @@ INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date
 ##      uniprot
 INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (1, 9031, 'ftp://ftp.ebi.ac.uk/pub/databases/SPproteomes/swissprot_files/proteomes/9031.SPC', '', now(), now(), "UniProtParser");
 
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 9031,'ftp://ftp.ncbi.nih.gov/genomes/Gallus_gallus/protein/protein.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
-
+##       refseq
+INSERT INTO source_url (source_id, species_id, url, checksum, file_modified_date, upload_date, parser) VALUES (3, 9031,'ftp://ftp.ncbi.nih.gov/genomes/Gallus_gallus/RNA/rna.gbk.gz', '', now(), now(), "RefSeqGPFFParser");
 
 ################################################################################