How to install local blast on linux

[yu_chem]

Hi everybody

Now I’m bothered with setting of local blast-2.2.26 on linux
Please help me. I'm beginner as bioinfomatician.

I have downloaded local blast-2.2.26, homo_sapiens genome and Refseq from below sites.
and use CentOS linux 7

genome data from ensembl ftp://ftp.ensembl.org/pub/release-75...o_sapiens/dna/
Refseq data from NCBI ftp://ftp.ncbi.nlm.nih.gov/genomes/H..._level.gff3.gz

Now I have two problems.
1.I don't know how to check proper combination of data between genome data and Refseq.
2.I don't know how to make database for blast using above two data

I think that actually user first combine two data into one file using 'cat' command and next use 'formatdb' command
>formatdb -i input-file-name -n database-name -p F

but I got following error
[formatdb] WARNING: Sequence number 1 (lcl|1_input-file-name), 8139970 illegal characters were removed:
54 Es, 114 Fs, 42 Is, 117 Js, 2 Ls, 114 Qs

Please reply to above question.
best regards

[Richard Finney]

That's a WARNING, not an ERROR.
Did it create an output file?

[GenoMax]

Why bother creating your own when NCBI makes pre-formatted indexes available here: ftp://ftp.ncbi.nlm.nih.gov/blast/db/ Both human genomic and refseq db are available.

Download the relevant files (there are multiple) for each set e.g. refseq_genomic.xx.tar.gz. Unzip/untar them and leave them all in a single folder. When doing the blast search use the "base name" as db name (e.g. refseq_genomic).

[yu_chem]

Dear Richard Finney.

Thank you for answer and I' sorry about delay of the reply
After I commanded formatdb, blast creates files: database-name.nsq, database-name.nhr and database-name.nin and then in formatdb.log, the below message was written.

Version 2.2.26 [Sep-21-2011]
Started database file "./data/Hchr-d.fa"
WARNING: [000.000] Sequence number 24 (lcl|24_Hdata), 8139970 illegal characters were removed:
54 Es, 114 Fs, 42 Is, 117 Js, 2 Ls, 114 Qs

Formatted 24 sequences in volume 0
SUCCESS: formatted database database-name

But e.g. when I commanded ./bin/fastacmd -d database-name -s NC_000001.10 -L 100,500, the blow message was displayed.

[fastacmd] ERROR: Entry "nc_000001.10" not found

If I succeed in formatdb, why cann't I succeed in fastacmd?

[yu_chem]

Dear Genomax

Thank you for answer and I'm sorry for delay of the reply.
I also try using other database.
It's reason for selecting these files of genome and genes that when I started to learn bioinformatics, other researcher recommended to use these files NCIB refseq and ensembl genome data, But I didn't know the reason. Actually I should ask him, But he is busy. So I try install of local blast on my own.