Making COGs, Running BLAST

[biolid]

Hi!
I am trying to assign COG domains to a multifasta protein file of a genome using COGnitor and I got stuck in the 3.1.0 Running BLAST step (which is indeed not yet part of the COgnitor procedure).

After command "makeblastdb" I obtained three output files: .phr, .pin and .psq.

After first "psiblast" (unfiltered BLAST results), I obtained a 35.8 MB .tab file

I am now performing second psiblast (filtered BLAST results), and a message appears like this: BLAST engine error: Traceback failed

I don't know what is happening because if the previous "psiblast" worked with the same files I don't know why it isn't working in this case.

It is the first time using BLAST and I am not a bioinformatics expert. I am doing this in a Biolinux virtual machine. I am using the BLAST installed by default.

Please, anyone can help me?

[biolid]

Ok, it was a mistake in the file. Now I cannot perform COGmakehash, it says that this command doesn't exist. I did "make" command to create the executable but it doesn't work. This is a little desperate . I have g++ compiler... If anyone knows something about this, please, help me

[GenoMax]

Did you get any errors when you tried "make" command?

[biolid]

Thank you for your interest. When I created the executables in COGmakehash and COGlse, no errors appeared. But when I try to create them in COGreadblast, COGtriangles and COGcognitor, I get this error:

g++ -O2 -c main.cpp
In file included from main.cpp:2:0:
os.h: In function ‘void myOpenDir(const char*)’:
os.h:13:16: error: ‘chdir’ was not declared in this scope
chdir(dirpath);
^
make: *** [main.o] Error 1


and the executables are not created

[GenoMax]

Try adding this include directive to main.cpp (see answer at this link): http://stackoverflow.com/questions/1...lation-error-g

[biolid]

main.cpp is only in COGcognitor folder, in other folders there are different .cpp files
-Folder COGlse: reader.cpp, mklse.cpp
-Folder COGmakehash: COGmakehash.cpp, ereader.cpp
-Folder COGreadbladst: cogreadblast.cpp, reader.cpp, bc.cpp, blastconv.cpp, reader.cpp
-Folder COGtriangles: cogmaker.cpp, cogtriangles.cpp

and where to include that line?:

#include "cognitor.h"
#include "os.h"
#include "cognitorglob.h"
#include "enum.h"
#include <unistd.h>

using namespace std;

int main(int argc, char* argv[]) // -p -d -o -a -c -n -t -e -m -g -w -l -r
{
double cogOverlapThreshold = 0.25;
double overlapThreshold = 0.5

Is it the right place?Sorry, I have no idea on this issues.

[biolid]

I included the line in .cpp file and the command make still not working. I will try in .h file

[GenoMax]

What version of g++/GCC are you using and on what OS?

Compilation worked for me on CentOS (gcc v.4.1.2) without editing any files.

[biolid]

gcc version 4.8.2

In Biolinux 8.0.3: Linux bl8vbox 3.13.0-32-generic #57-Ubuntu SMP Tue Jul 15 03:51:08 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

[GenoMax]

Add this line

#include "unistd.h"

to top of os.h file in COGcoginitor/COGtriangles, bc.h in COGreadblast. Needed this for gcc v. 4.8.1.

Remove it from main.cpp if you had added it there.

[biolid]

I did what you said and the executables were created but, after that (I think) the command doesn-t work:

manager@bl8vbox[COGmakehash] COGmakehash -h
zsh: command not found: COGmakehash

And this happens in all the cases.What am I doing wrong?

[GenoMax]

You need to do

$ ./COGmakehash

Since the directory you are in is not in your $PATH.

[biolid]

Oh, thank you so much, I still don't understand well the function of commands. I really thank you for your help

[biolid]

Hi again! I have some doubts again:

- One thing I dont understand is the first step in all the process and the format files. There are some steps in *Procedure* in READme file, but I don-t know if you have to do the part 3.1 or just 3.2.

- A second thing is that I have multifasta file with aa sequences of a genome with this entries (as examples):
>locustag_0001 hypothethical protein
>locustag_0002 TelA-like protein
....

I wonder how could it be the file GenQuery.p2o.csv with format <protein-id>,<genome-id>:
0001,locustag
0002,locustag
or
hypothethical protein, locustag
TelA-like protein,locustag??

- Is COG.p2o.csv the same as cognames2003-2014.tab without second column or cog2003-2014.csv?

- Which is the file COGs.csv?In README file says it is tha output of COGtriangles program, but I dont know even how to perform COGlse because I have only one genome...Otherwise, I read in a forum that this file is cog2003-2014.csv, which one then?

Thank you in advance.