Error while makeblastdb

Question

[sp]

I downloaded human chromosome from UCSC and tried to make a "blastable" database. Each chromosomes was saved separately, something like chr1.fa, chr2.fa etc. I would like to make one "BIG" and "blastable" database file named as "humna38.fa.

I wrote a command like "makeblastdb -in ~/Database/Chr38/*.fa -dbtype nucl -out human38.fa", but gave me error "Too many positional arguments (1), the offending value: chr10.fa"

What is wrong and fix?

[####]

Use following steps :

1.First ,concatenate all the chromosome sequences into one using cat function of linux 2.Use following command for making database:

makeblastdb -in *.fa -input_type fasta -dbtype nucl -title xxx -out human_db

make sure fasta sequence is in same directory or else give the complete path.

I hope this will help

[sp]

Should I concatenate first by cat *.fa > new_name.fa? and then makeblastdb -in new_name.fa -input_type fasta -dbtype nucl -out human_db?

[5heikki]

Yes, the error message is relatively clear about that. "Too many positional arguments (1), the offending value: chr10.fa"

[####]

yes u have to concatenate first

Answer 1

[onspotproductions]

Why don't you simply use the remote search function in Blast+? If not then you could try concatenating the .fa files into a single large file and using makeblastdb, but I am not sure if it will like a file of that size.

[sp]

I would like to make index file for "bwa". That's why.

Answer 2

[Jordan Anaya]

I'm not sure that makeblastdb can take multiple files at once.

[sp]

Thanks for your comment. If it's true, what will be the way to make one big database file of human genome? In theory, all the chr seq can be consolidated and makeblastdb it after. I do not know how though. Can you please correct me if I am wrong?

[Jordan Anaya]

Hmm, I've never made a database with multiple files before. This previous post on biostars is old but hopefully is still relevant: Formatdb From A Batch Of Fasta Sequences

It seems that you can make a database for each chromosome separately, then use the blastdb_aliastool to make an alias for the collection which can then be searched against.