Blast database creation : Error

[kalyankpy]

HI,

I am new to Bioinformatics. I am creating a database with the fasta formatted sequences from high througput sequncing data (20-30 bp in length). While I create the blast database (version ncbi-blast-2.2.23), I notice an error which is not comprehendible to me. I request you to help me:

Command used:
============
$ makeblastdb -in puts2.fa -dbtype nucl -parse_seqids -mask_data puts2_mask.asnb -out puts2db -title "puts2_database"

Error message:
===========
Error: (803.7) Bad char in VisibleString: 9

Do help me in solving this issue

[bcanback]

makeblastdb Bad char in VisibleString

Hi!

I had the same message today. In my case it is was due to a tab found in the id line. Hope your problem is equally simple.

cheers

Björn

[kalyankpy]

Hi Bjorn,

In my case there wasnt any problem in the data. Rather when I uninstalled the ....24 version of the blast++ and worked with 23 version things worked out pretty smooth. Later I noticed that there is a bug in that version.

[fishinabarrel]

Had the same issue.

Tab was causing the problem in the header.


Simple fix in vi.
From the colon :

:%s/\t//g

*replaces all tabs with nothing.

[kaem]

Hi all

I also create with makeblastdb 2.2.26+ - I received the same error message+
Error: (803.7) Bad char in VisibleString: 9
and deleted the tabs.

however, making a database from Aedes Aedes_aegypti_Aaeg_1.2 i had two other error messages:

Error: (803.7) Bad char in VisibleString: 195
Error: (803.7) Bad char in VisibleString: 164

Does anybody know what it means?
thanks kaem

[PenelopeFudd]

Apparently the number at the end of the message (9, 164, or 195) is the character code that it didn't like. So, the '9' is a tab character, and 164 and 195 are just random characters that together probably mean 'dash' in Unicode or something.

To fix it, edit the fasta file with 'vi', search for all characters that aren't printable, and delete them or something. I used '/[^ -|]/' (match anything outside the range of space to vertical-bar) to find the bad character in my file. It was between two words in a header line, not in a sequence: "thioredoxin(?)peroxidase", so I replaced it with a dash.

Hope this helps you, or anyone googling after you!

Hi all

I also create with makeblastdb 2.2.26+ - I received the same error message+
Error: (803.7) Bad char in VisibleString: 9
and deleted the tabs.

however, making a database from Aedes Aedes_aegypti_Aaeg_1.2 i had two other error messages:

Error: (803.7) Bad char in VisibleString: 195
Error: (803.7) Bad char in VisibleString: 164

Does anybody know what it means?
thanks kaem

[Benoit_UMons]

Hi all,

How works the software with this error? They correct it?

Because, my database is too big (13Go). I can't open it to correct the error manually.

My bad character is 0X92 (i thing it's " , ").

Can you help me please?

Thank you very much,

Benoit

[jjzieve]

I'm new to seqanswers so forgive me for not knowing how to start a new thread. I'm having a strange issue... I'm running "makeblastdb -in seq.fsa -dbtye nucl" everything finishes fine (no error message) but there is no output (i.e. *.nin, *.nhr, *.nsq). I want to keep the outputs the same name as the inputs so I didn't specify "-out". I figured it might have saved in the $BLASTDB location but that isn't even set. One note is the fasta file is 16 GB, do I need to manually split it up and create *.nal or something? Thanks

[GenoMax]

I'm new to seqanswers so forgive me for not knowing how to start a new thread. I'm having a strange issue... I'm running "makeblastdb -in seq.fsa -dbtye nucl" everything finishes fine (no error message) but there is no output (i.e. *.nin, *.nhr, *.nsq). I want to keep the outputs the same name as the inputs so I didn't specify "-out". I figured it might have saved in the $BLASTDB location but that isn't even set. One note is the fasta file is 16 GB, do I need to manually split it up and create *.nal or something? Thanks

Starting new threads:

SeqAnswers.com --> "Forums" (from site navigation at top left) --> Choose an appropriate forum --> "New Thread" button at top left.

It is always a good idea to search the forum first with some keywords.

See this thread: http://seqanswers.com/forums/showthread.php?t=34619 and the link in the thread to NCBI blast command line manual.

How much memory do you have on this machine? For a 16GB sequence file it should take a while to make the database. How long did your job take?

[jjzieve]

We have 16 gigs of RAM, the job only took about 5 minutes. So... I'm thinking maybe the job quit with no errors? I have access to a high memory machine on a HPC I'll try running it there and let you know how that works out.

[GenoMax]

In case you still have trouble try to capture the standard error/out messages to a file to see if that provides any additional info.

[jjzieve]

Worked like a charm on a high memory node running ubuntu 13.04. Do you think it would be worthwhile to put in a bug request to the blast people? That essentially, their were no errors reported for what was probably a stack overflow? The other system I was on was CentOS 6.4, btw.