Menu

#14 mpiformatdb creates illegal databases

open
2
2004-11-30
2004-10-03
Lucas Carey
No

mpiformatdb creates databases with duplicate gi's

corect behavior:
engrailed {117} cp yeast.nt yeast
engrailed {118} cat yeast.nt >> yeast

engrailed {125} formatdb -o T -V T -p F -i yeast
echo $?[formatdb] ERROR: Failed to create index. Possibly a gi
included more than once in the database.

engrailed {126} echo $?
9

engrailed {127} formatdb -o T -p F -i yeast
[formatdb] ERROR: Failed to create index. Possibly a gi included
more than once in the database.

engrailed {128} formatdb -p F -i yeast
engrailed {129} echo $?
0
engrailed {130} formatdb -p F -V T -i yeast
engrailed {131} echo $?
0

mpiformatdb:
engrailed {350} ./mpiformatdb -N 2 -i ~/Blast/DBs/yeast -V T
Reading input file
Done, read 303926 lines
Reordering 34 sequence entries
Database type unspecified, assuming nucleotide
Trying to break yeast (25 MB) into 2 fragments of 13 MB
Executing: formatdb -V T -i /var/tmp/reorderncgBpq -p F -N 2 -n /
users/lcarey/Blast/Shared//yeast -o T
Created 2 fragments.

Discussion

  • Aaron Darling

    Aaron Darling - 2004-10-03

    Logged In: YES
    user_id=756995

    Looks like this is a bug in formatdb as well as mpiformatdb:

    koadman@narmer:/home/mpiblast>
    ~/software/ncbi-062004/build/formatdb -i double.nt -o T -p F
    [formatdb] ERROR: Failed to create index. Possibly a gi
    included more than once in the database.

    koadman@narmer:/home/mpiblast>
    ~/software/ncbi-062004/build/formatdb -i double.nt -o T -p F
    -v
    koadman@narmer:/home/mpiblast> du -sh double.nt
    24M double.nt
    koadman@narmer:/home/mpiblast>
    ~/software/ncbi-062004/build/formatdb -i double.nt -o T -p F
    -v 3
    koadman@narmer:/home/mpiblast> cat double.nt.nal
    #
    # Alias file created Sun Oct 3 18:15:55 2004
    #
    #
    TITLE double.nt
    #
    DBLIST double.nt.00 double.nt.01 double.nt.02 double.nt.03
    double.nt.04 double.nt.05 double.nt.06 double.nt.07 double.nt.
    08
    #
    #GILIST
    #
    #OIDLIST
    #

    So long as the duplicate gid's end up in different fragments it
    doesn't know any better.

     
  • Aaron Darling

    Aaron Darling - 2004-11-30
    • priority: 5 --> 2
     

Log in to post a comment.

MongoDB Logo MongoDB