wrong version of formatdb was used to make database

Question

[shalinikaushik1293]

I am using uniref100 database for running the psipred script (runpsipred) on macOS. On running the runpsipred script :

#!/bin/tcsh

# This is a simple script which will carry out all of the basic steps
# required to make a PSIPRED prediction. Note that it assumes that the
# following programs are in the appropriate directories:
# blastpgp - PSIBLAST executable (from NCBI toolkit)
# makemat - IMPALA utility (from NCBI toolkit)
# psipred - PSIPRED V4 program
# psipass2 - PSIPRED V4 program

# NOTE: Script modified to be more cluster friendly (DTJ April 2008)

# The name of the BLAST data bank
set dbname = /Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta

# Where the NCBI programs have been installed
set ncbidir = /Users/shalini/Desktop/pgen_github/ncbi-blast-2.10.0+/bin

# Where the PSIPRED V4 programs have been installed
set execdir = /Users/shalini/Desktop/pgen_github/pGenTHREADER-master/psipred/bin

# Where the PSIPRED V4 data files have been installed
set datadir = /Users/shalini/Desktop/pgen_github/pGenTHREADER-master/psipred/data

set basename = $1:r
set rootname = $basename:t

# Generate a "unique" temporary filename root
set hostid = `hostid`
set tmproot = psitmp$$$hostid

\cp -f $1 $tmproot.fasta

echo "Running PSI-BLAST with sequence" $1 "..."

$ncbidir/blastpgp -b 0 -v 5000 -j 3 -h 0.001 -d $dbname -i $tmproot.fasta -C $tmproot.chk >& $tmproot.blast

if ($status != 0) then
    tail $tmproot.blast
    echo "FATAL: Error whilst running blastpgp - script terminated!"
    exit $status
endif

echo "Predicting secondary structure..."

echo $tmproot.chk > $tmproot.pn
echo $tmproot.fasta > $tmproot.sn

$ncbidir/makemat -P $tmproot

if ($status != 0) then
    echo "FATAL: Error whilst running makemat - script terminated!"
    exit $status
endif

echo Pass1 ...

$execdir/psipred $tmproot.mtx $datadir/weights.dat $datadir/weights.dat2 $datadir/weights.dat3 > $rootname.ss

if ($status != 0) then
    echo "FATAL: Error whilst running psipred - script terminated!"
    exit $status
endif

echo Pass2 ...

$execdir/psipass2 $datadir/weights_p2.dat 1 1.0 1.0 $rootname.ss2 $rootname.ss > $rootname.horiz

if ($status != 0) then
    echo "FATAL: Error whilst running psipass2 - script terminated!"
    exit $status
endif

# Remove temporary files

echo Cleaning up ...
\rm -f $tmproot.* error.log

echo "Final output files:" $rootname.ss2 $rootname.horiz
echo "Finished."

The error is

Running PSI-BLAST with sequence /Users/shalini/Desktop/pgen_github/unmodelled_fasta/NP_999840.fasta ...
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.65.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.66.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.67.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.68.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.69.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.70.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.71.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.72.
[NULL_Caption] WARNING: readdb: wrong version of formatdb was used to make database //Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta.73.
[NULL_Caption] WARNING: Unable to open uniref100.fasta.pin
FATAL: Error whilst running blastpgp - script terminated!

I have made the database using

makeblastdb -in /Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta -parse_seqids -blastdb_version 5 -title "Unirefdb" -dbtype prot

results in :

Building a new DB, current time: 01/14/2020 15:45:24
New DB name:   /Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta
New DB title:  Unirefdb
Sequence type: Protein
Keep MBits: T
Maximum file size: 1000000000B
Adding sequences from FASTA; added 199397329 sequences in 7009.84 seconds.

There are total 73 .psq, .pin, .pog and .phr files generated

[lieven.sterck]

where did you get the blastpgp executable from? ( as far as I know it's not part of the default blast package) and consequently which version are you using?

perhaps try formatting your DB using the older -blastdb_version 4 . In any case when formatting large DB you should also get one .pal (or .nal for nucleotides) files which 'groups' all the subparts.

Answer 1

[Mensur Dlakic]

Most likely you need to format your protein database with formatdb that matches your version of blastpgp. Based on your blast programs location, you format the database like this:

/Users/shalini/Desktop/pgen_github/ncbi-blast-2.10.0+/bin/formatdb -i /Users/shalini/Desktop/pgen_github/uniref100/uniref100.fasta -p T -t "Unirefdb"

You may want to first delete previously formatted files.

[lieven.sterck]

formatdb is been deprecated for years now, so I don't think it will be available from the 2.10.0+ version

[Mensur Dlakic]

Right you are about formatdb being deprecated. On the other hand, PSI-PRED still works best with legacy BLAST versions that have formatdb, so here is the link to last version that should work:

ftp://ftp.ncbi.nlm.nih.gov/blast/executables/legacy.NOTSUPPORTED/2.2.26/

[shalinikaushik1293]

Thank you so much for your reply.

I resolved the problem by using formatdb. But after running the program for 3 hours for a single protein. It gives me the error:

Running PSI-BLAST with sequence /Users/shalini/Desktop/pgen_github/unmodelled_fasta/AAA18895.fasta ...

[blastpgp 2.2.26] WARNING:  [000.000]  Failed to access string index ISAM Error code is -5

[blastpgp 2.2.26] WARNING:  [000.000]  Failed to retrieve sequence lcl|UniRef100_T1G0H5
[blastpgp 2.2.26] WARNING:  [000.000]  Failed to access string index ISAM Error code is -5

[blastpgp 2.2.26] WARNING:  [000.000]  Failed to retrieve sequence lcl|UniRef100_T1G0H5
blastpgp(1042,0xa986a1c0) malloc: *** mach_vm_map(size=8388608) failed (error code=3)
*** error: can't allocate region
*** set a breakpoint in malloc_error_break to debug
[blastpgp 2.2.26] FATAL ERROR: CoreLib [001.000]  Failed to allocate 40100 bytes
FATAL: Error whilst running blastpgp - script terminated!

[shalinikaushik1293]

I am using macOS High Sierra, processor is 3.3 GHz Intel Core i7 and the memory is 16 GB 1867 MHz DDR3.

Looking forward for your reply. I need your help.

[Mensur Dlakic]

I am not an expert on BLAST error messages, but it could be that you don't have enough memory. I suggest you try UniRef90 database as it is considerably smaller.