New member with a blast database problem

[rbruenn]

Hello!

I am a graduate student at UC Berkeley currently working with raw reads of several transcriptomes in an attempt to find and assemble reads that match a couple of genes I'm studying. This site has already been very useful to me (Thank you!!!!), but I haven't found any answers pertaining to my current problem.

I was hoping I could get some help with a BLAST problem I'm having. I am working with standalone blast, and am building blast databases from fasta files of transcriptome raw reads. I have successfully used the command:

makeblastdb -in COST_1_final.fasta -input_type fasta -dbtype nucl -out COST_1_final for several fasta files of raw reads (with varying names of course) but a few of the files result in multiple sets of database files, marked for example <filename>.00.nhr and <filename>.01.nhr with what I believe is an alias file <filename>.nal

The command as it runs give the usual message:
Building a new DB, current time: 10/15/2015 10:49:55
New DB name: COST_u_final
New DB title: COST_u_final.fasta
Sequence type: Nucleotide
Keep Linkouts: T
Keep MBits: T
Maximum file size: 1000000000B
Adding sequences from FASTA; added 16313457 sequences in 937.429 seconds.

but then results in multiple sets of files.

Any ideas about how I can make just one set of database files, as has successfully happened with the rest of my fasta files? I would really appreciate any help!!

[GenoMax]

It is normal to get multiple files per blast database. That is how makeblastdb is supposed to work. Just make sure files for a database stay together in the same directory and you use the "basename" for the database (a suggestion: name your database some thing other than your input file name) when you run your searches.

[rbruenn]

I've always gotten multiple files in the sense of .nhr, .nin, .nsq files, but I am getting 2 of each, like .00.nhr and .01.nhr. Why would that only happen some of the time?

[GenoMax]

That is probably dependent of the size of the input fasta file. e.g. nt database has 32 fragment files now.

There should a database_name.nal file that enumerates all the file pieces if there are more than one.

[rbruenn]

Ahhh, interesting! So if I search using the command

blastn -db COST_u_final.fasta -query Genefiles -outfmt 6 -out BLASTresults.txt

will it search all of the files that were made?

Thank you for your help! I had no idea this wasn't a problem since I've never seen this happen before.

[GenoMax]

Yes that is correct.

Just to make things less confusing (to others, if needed later on) don't use the fasta file name as the -out database basename.

[rbruenn]

Thank you for the advice, I will change that practice.

[westerman]

Hum. I often use the fast file name as the blastDB name. Keeps them together. The extensions are going to be different so there should be no confusion. @GenoMax: what do you use?

[GenoMax]

I generally drop the .fasta/.fa part when naming a blast db. (Like NCBI. They don't call their db's nt.fa or nr.fa).

The idea of using just the "basename" when specifying a db index is a new for some. I suppose keeping the .fasta/.fa may be more logical for them.