[solved] Retrieve fasta from balst db using blastdbcmd: Error: gi|742519789: OID not found

Question

[dago]

I saw that many people had problem retrieving sequences from blast db.

I could find a way around it, so maybe someone has a god link/suggestion/reference I could use.

I want to extract sequences from nr db.

I have a list of identifier, obtained from a previous blast search

gi|740719731|ref|WP_038505017.1|
gi|740813732|ref|WP_038599015.1|
gi|740864652|ref|WP_038649903.1|
gi|740899195|ref|WP_038684443.1|
gi|740906294|ref|WP_038691542.1|

Now I try to query only

GIs:

740864652
740899195
740906294

or ref:

WP_038649903.1
WP_038684443.1
WP_038691542.1

blastdbcmd -db ~/Documents/nr_blastdb/nr -entry_batch Ids

But I get always:

Error: XXXXX: OID not found

What am I missing here?

Answer 1

[5heikki]

Did you make the nr db yourself from the fasta file or download the pre-formatted db files? If former, did you apply the -parse_seqids flag? If not, there's your problem.

[dago]

Not quite sure. I got it from a colleague. Could it be a problem related to the index of the entries?

[5heikki]

What does cat ~/Documents/nr_blastdb/nr.pal show?

If it's a pre-formatted db, the title line is something like:

TITLE All non-redundant GenBank CDS translations+PDB+SwissProt+PIR+PRF excluding environmental samples from WGS projects

[dago]

Perfect, was a manually created db. I will try to follow your first suggestion

[dago]

You are right, it worked fine now. However, for some seq now there are really crazy Ids, for example

> XXXXXX >XXXXX
CGSNDHIEJWPSP

It looks like two Ids one after the other. Any idea where the problem is?

[5heikki]

It's a non-redundant database and those two accessions encode an identical protein. You can avoid this behavior with the -target_only flag..

[dago]

I learned many things today! Thanks very much!!!!