Ensembl database fetch sequence problem

[roger770507]

hello
i just download the mySQL data from Ensembl
and successful construct the database (homo spaien core 73)
but my problem is if i had a ENST ID (ex: ENST00000575820)
how do i fetch the sequence via the database?
i already read the manual but haven't found any info how to do it.
plz help me , thanks!

[ffinkernagel]

It's not straight forward, since you'll need to retrieve all the exons and join them by hand.
Here's a bit of Python + SQL that yields all transcript_id->cDNA pairs - I'm sure you can adapt it (works for any revision >= 65)
You'll also need your own reverse complement function...

      cur.execute("""SELECT 
                        exon.seq_region_id, 
                        exon.seq_region_start, 
                        exon.seq_region_end, 
                        exon.seq_region_strand,
                        transcript.stable_id
                        FROM
                        exon, exon_transcript, transcript, seq_region
                        WHERE
                        exon.exon_id = exon_transcript.exon_id AND
                        transcript.transcript_id = exon_transcript.transcript_id 
                        AND seq_region.seq_region_id = exon.seq_region_id
                        AND seq_region.seq_region_id = %s
                        ORDER by exon_transcript.transcript_id, exon.seq_region_start
                        """, ( seq_region_id,))
            chr_seq = self.get_chromosome_sequence(conn, seq_region_id, seq_region_length)

            for transcript_id, rows in itertools.groupby(cur.fetchall(), lambda x: x[4]):
                rows = list(rows)
                strand = rows[0][3] 
                seqs = []
                #print transcript_id, rows
                for row in rows:
                    seqs.append(chr_seq[row[1] - 1:row[2]]) #ensembl intervals are 1 based, inclusive.
                if strand == 1:
                    seq = "".join(seqs)
                else:
                    seq = common.reverse_complement("".join(seqs)).upper()
                found = True
                yield (transcript_id, seq)

Best regards,
F

[Giulietta EnsemblHelpdesk]

We suggest the Ensembl Core Perl API or REST API for getting sequences from genes and transcripts. The MySQL query as you see is inefficient and involved, and the APIs are designed for fast and simpler querying:

http://www.ensembl.org/info/docs/api/core/index.html#api

http://beta.rest.ensembl.org/