[maubp]

This version looks for the first copy of 'blastn' in the list of folders defined with $PATH,

blastn -query /home/viraja/Desktop/samples_with_no_flags/sample_01/contigs_01_without_flag.fasta -db refseq -num_alignments 1 -outfmt '6 qseqid stitle'

By specifying the path to BLAST 2.2.28+, you tell Linux exactly where to look. We mean sdo omething like this, where I'm assuming you downloaded and decompressed the BLAST 2.2.28+ files under /home/viraja/downloads/ - adjust to match your setup:

/home/viraja/downloads/ncbi-blast-2.2.28+/bin/blastn -query /home/viraja/Desktop/samples_with_no_flags/sample_01/contigs_01_without_flag.fasta -db refseq -num_alignments 1 -outfmt '6 qseqid stitle'

[rhinoceros]

I totally agree with you kmcarr. Its very confusing for me too. I got a brand new computer and I installed everything myself. So I know whats on the computer.

I bet you downloaded and uncompressed blast 2.2.28 and then left it at that (didn't put it into your path), went to the terminal and typed blastn and then your distro pulled blast 2.2.27 from some repo in which the latest blast is 2.2.27.

[vpi]

I downloaded it to the downloads folder, but then I moved it to home directcory and decompressed it there.

[rhinoceros]

I downloaded it to the downloads folder, but then I moved it to home directcory and decompressed it there.

The general way to put things into your path is to add them to .bashrc or .bash_profile (or whatever your distro has). In terminal type

cd

and then

ls -la

and you'll see some file like this. Next edit it with for example nano, i.e.

nano .bashrc

and add the path there, e.g.

export PATH=/Users/you/Software/ncbi-blast-2.2.28+/bin:${PATH}

Next

source .bashrc

and if you added the correct bin directory (what you see when you type pwd in the blast dir), it's now in your path. It's sort of like creating shortcuts in Windows.

[vpi]

Both of you were correct.

Now after chnaging the path,

echo $PATH
/usr/lib/lightdm/lightdm:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/home/me/ncbi-blast-2.2.28+/bin


blastn -version
blastn: 2.2.27+
Package: blast 2.2.27, build Oct 29 2012 07:36:27

But still says blast 2.2.27+

[kmcarr]

Both of you were correct.

Now after chnaging the path,

echo $PATH
/usr/lib/lightdm/lightdm:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/home/me/ncbi-blast-2.2.28+/bin


blastn -version
blastn: 2.2.27+
Package: blast 2.2.27, build Oct 29 2012 07:36:27

But still says blast 2.2.27+

Because "/usr/bin" which is where 2.2.27+ is located, comes BEFORE "/home/me/ncbi-blast-2.2.28+/bin" in your $PATH variable.

When you have more than one copy of program installed, *NIX systems choose the one which comes FIRST in your $PATH. You have to rearrange the order of directories in your $PATH.

[rhinoceros]

Like kmcarr said except instead of rearranging $PATH you could just delete the blast 2.2.27+ binaries from "/usr/bin".

[vpi]

Could you please explain how can I delete them. I tried to delete them directly from the folder but I failed.

[rhinoceros]

Could you please explain how can I delete them. I tried to delete them directly from the folder but I failed.

rm /usr/bin/blastn

Removes the blastn binary. If it says something about rights then:

sudo rm /usr/bin/blastn

There will probably still be other blast 2.2.27+ related binaries in your "/usr/bin". If you installed it by apt-get, this describes how to remove it.

[vpi]

Thank you so much. Everything works fine now. But I have a new issue.

The blast output file has repeated results, I mean the query, subject is repeated several times, even though I use -num_alignments 1. For example,

NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52

[rhinoceros]

Thank you so much. Everything works fine now. But I have a new issue.

The blast output file has repeated results, I mean the query, subject is repeated several times, even though I use -num_alignments 1. For example,

NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52

As I wrote earlier, -num_alignments (Show alignments for this number of database sequences) is an invalid flag for outfmt 6 since it doesn't show any alignments to begin with. There are other flags that you could try in the blast manual. This thread is also highly relevant.

[kmcarr]

Thank you so much. Everything works fine now. But I have a new issue.

The blast output file has repeated results, I mean the query, subject is repeated several times, even though I use -num_alignments 1. For example,

NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52
NODE_20_length_864_cov_1.18725_ID_1126030 gi|219666071|ref|NC_011830.1| Desulfitobacterium hafniense DCB-2, complete genome 2e-52

'-num_alignments' only applies to output formats which actually include alignments (e.g. outfmt 0,1,2,3,4); it is meaningless (and therefore ignored) for tabular output formats like yours.

Parameters which limit the number of reported hits in tabular output are -max_target_seqs and -max_hsps_per_subject.

But you should be aware that by limiting the number of alignments reported for each query you may be throwing away important data. What the output you have shown above indicates is that this particular query (NODE_20) aligns to multiple locations on the genome of Desulfitobacterium hafniense DCB-2. That may be important!