Extracting information from EMBL flat file

[valente]

Hey guys,
I have a file of the proteome set of C. elegans that I retrieved from Uniprot in a EMBL flat file like this:

EMBL_FLAT_FILE_CELEGANS

ID   14331_CAEEL             Reviewed;         248 AA.
AC   P41932; Q21537;
DT   01-NOV-1995, integrated into UniProtKB/Swiss-Prot.
DT   22-JUL-2008, sequence version 2.
DT   28-NOV-2012, entry version 95.
DE   RecName: Full=14-3-3-like protein 1;
DE   AltName: Full=Partitioning defective protein 5;
GN   Name=par-5; Synonyms=ftt-1; ORFNames=M117.2;
OS   Caenorhabditis elegans.
OC   Eukaryota; Metazoa; Ecdysozoa; Nematoda; Chromadorea; Rhabditida;
OC   Rhabditoidea; Rhabditidae; Peloderinae; Caenorhabditis.
DR   GO; GO:0005938; C:cell cortex; IDA:WormBase.
DR   GO; GO:0005634; C:nucleus; IDA:WormBase.
DR   GO; GO:0045167; P:asymmetric protein localization involved in cell fate determination; IMP:WormBase.
DR   GO; GO:0001708; P:cell fate specification; IMP:WormBase.
DR   GO; GO:0043053; P:dauer entry; IMP:WormBase.
DR   GO; GO:0008340; P:determination of adult lifespan; IMP:WormBase.
DR   GO; GO:0009792; P:embryo development ending in birth or egg hatching; IMP:WormBase.
DR   GO; GO:0000132; P:establishment of mitotic spindle orientation; IMP:WormBase.
DR   GO; GO:0030590; P:first cell cycle pseudocleavage; IMP:WormBase.
DR   GO; GO:0035188; P:hatching; IMP:WormBase.
DR   GO; GO:0007126; P:meiosis; IMP:WormBase.
DR   GO; GO:0002009; P:morphogenesis of an epithelium; IMP:WormBase.
DR   GO; GO:0009949; P:polarity specification of anterior/posterior axis; IMP:WormBase.
DR   GO; GO:0035046; P:pronuclear migration; IMP:WormBase.
DR   GO; GO:0006898; P:receptor-mediated endocytosis; IMP:WormBase.
DR   GO; GO:0007346; P:regulation of mitotic cell cycle; IMP:WormBase.
DR   GO; GO:0010070; P:zygote asymmetric cell division; IMP:WormBase.
SQ   SEQUENCE   248 AA;  28191 MW;  ABBE0DA27D9341AF CRC64;
     MSDTVEELVQ RAKLAEQAER YDDMAAAMKK VTEQGQELSN EERNLLSVAY KNVVGARRSS
     WRVISSIEQK TEGSEKKQQL AKEYRVKVEQ ELNDICQDVL KLLDEFLIVK AGAAESKVFY
     LKMKGDYYRY LAEVASEDRA AVVEKSQKAY QEALDIAKDK MQPTHPIRLG LALNFSVFYY
     EILNTPEHAC QLAKQAFDDA IAELDTLNED SYKDSTLIMQ LLRDNLTLWT SDVGAEDQEQ
     EGNQEAGN
//

NOTE: the file showed is here shortened.


Moreover, I have another file with a lot of gene full names and I would like to extract informations of GO for these genes from the EMBL flat file. In other words, I would like to know if someone here have some script that read my file with the gene full names (one per line), find it in this EMBL flat file and extract the GO. The output desirable is the gene full name followed by its gene ontology separated by comma (including each ontology).

OUTPUT

GENE_A,GO; GO:0001708; P:cell fate specification; IMP:WormBase, GO; GO:0043053; P:dauer entry; IMP:WormBase,GO; GO:0008340; P:determination of adult lifespan; IMP:WormBase,GO; GO:0009792; P:embryo development ending in birth or egg hatching; IMP:WormBase, GO; GO:0000132; P:establishment of mitotic spindle orientation; IMP:WormBase

If you guys have other ideas it would be nice!

Cheers.

[maubp]

Do you know any scripting/programming language? Both BioPerl and Biopython (and likely other libraries too) could assist you with their EMBL parsers - although in this case you could do this without a full parser.

[jiaco]

biomaRt (R/bioconductor): http://www.bioconductor.org/packages...l/biomaRt.html

library( biomaRt )

uniprot = useMart( "unimart" );
uniprot = useDataset( "uniprot", uniprot );

# these can be looked at for more options in search(filters) and retrieve(attributes)

filters = listFilters( uniprot );
attributes = listAttributes( uniprot )

useFilter = c( "accession" );
useAttributes = c( "accession", "gene_name", "go_id", "go_name" );

query = "P41932";
df = getBM( mart=uniprot, values=c(query), filters=useFilter, attributes=useAttributes )

nrow = dim( df )[ 1 ];
s=sprintf( "%s", df[1,2] );
for( i in 1:nrow ) {
        s = sprintf( "%s,GO; %s; %s;", s, df[i,3], df[i,4] );
}

If you have a text file full of accessions and want output with 1 gene per line:

query = read.table( "queryfile.txt" );
# assume 1st column is accession

query = as.character( query[,1] );

mdf = getBM( mart=uniprot, values=query, filters=useFilter, attributes=useAttributes )

uniqueAccs = unique( sort( as.character( mdf[,1] ) ) );
outvec = vector( mode="character", length=0 );
for( acc in uniqueAccs ) {
        df = mdf[ mdf[,1] == acc, ];
        nrow = dim( df )[ 1 ];
        s=sprintf( "%s", df[1,2] );
        for( i in 1:nrow ) {
                s = sprintf( "%s,GO; %s; %s;", s, df[i,3], df[i,4] );
        }
        outvec = c( outvec, s );
}
write.table( outvec, "myoutfile.txt", quote=F, row.names=F, col.names=F );

(the second code snippet depends on the preamble from the first)

EDIT: I realize I did not answer your question, but this will get the job done without any need for downloading embl files.