Make_fragments.pl database

[Anonymous]

Anyone knows the right database to download from the ncbi site?
The link in the rosetta tutorial is broken. I know that’s the nr one, but fasta formated or env.nr..
any input on this?

thank you

tiago

[Anonymous]

> Anyone knows the right database to download from the ncbi site?
> The link in the rosetta tutorial is broken. I know that’s the nr one, but fasta formated or env.nr..
> any input on this?
>
> thank you
>
> tiago

You can use UniRef90 from http://www.uniprot.org/. Fasta formated and you a have to run formatdb for setting up the psiblast database files. Just rename the UniRef90 fasta to nr and run

‘/my/blast/installation/formatdb -o T -i nr’

David K

[Anonymous]

ok

thanks ive had done and it worked, something to do with the index file, now it says it doesnt find the blosum65.iqj matrix, but its in the right directory. i just copied nnmake_database to my directory. does the script make_fragments.pl need more data from the fragments directory in rosetta.

[Anonymous]

Hi Tiago

Where did you download nnmake_database?

In the tutorial page it says provided with the release but I dont see anywhere?

I have difficulty to modify the make_fragment.pl script as I dont find following database and ss prediction software sam?

# nnmake

my $VALL = "$scratch/$fragsDir/nnmake_database";          # dir containing vall database (cvs repository 'nnmake_database')

my $VALL2 = "$scratch/$fragsDir/nnmake_database";             # alt dir containing vall database (cvs repository 'nnmake_database')

my $VALL_NAME = "vall.dat.2006-05-05";                 # filename of vall (vall.dat.<id> and vall_cst_coord.dat.<id> must exist)

my $NNMAKE = $pathToDemo."scripts/frags/make.sh";                         # nnmake binary  (cvs respository 'nnmake')

my $TRIMLOOPS = "$scratch/$fragsDir/nnmake/trimLoopLibrary.pl";                  # trimLoopLibrary.pl (cvs respository 'nnmake')



# chemshift

my $CHEMSHIFT = "$scratch/$fragsDir/chemshift/pCHEMSHIFT.gnu";     # chemshift binary (cvs repository 'chemshift')

my $TALOS_DB = "$scratch/$fragsDir/chemshift_database";                          # TALOS databases directory (cvs respository 'chemshift_database')



# jufo (secondary structure prediction software)

my $JUFO = "$scratch/src/jufo/molecule.exe";                              # jufo executable



# sam (secondary structure prediction software)

my $SAM_target99                = "$scratch/src/sam3.5.i686-linux/bin/target99";        # sam target99 executable

my $SAM_uniqueseq               = "$scratch/src/sam3.5.i686-linux/bin/uniqueseq";       # sam uniqueseq executable

my $SAM_predict_2nd_dir         = "$scratch/src/sam.predict-2nd/";                      # sam predict-2nd directory

my $SAM_predict_2nd             = "$scratch/src/sam.predict-2nd/predict-2nd";   # sam predict-2nd executable

Can anyone help me running make_fragment.pl?

I also dont know what is the directory name: $fragsDir = “rosetta/rosetta_fragments_latest”;

as I dont know anything about perl script does this means we create a directory and give path like this?

 

kind regards

Mustafa

[Anonymous]

 

Hi

I am trying to use following link to generate fragments

https://www.rosettacommons.org/manuals/archive/rosetta3.1_user_guide/file_fragments.html

but I cant find the nnmake_database given in the link. Can anyone share the correct link? and also help in running the make_fragment.pl as it fails to run due to failure of nr database download. However, if I select uniref96 it downloads correctly.

kind regards

Mustafa

[Anonymous]

I am not familar with the Perl script described or fragments stuff in general, but I thought I’d mention that one can do this easily with PyRosetta too.

pose = pyrosetta.pose_from_file('****.pdb')

confragset = pyrosetta.rosetta.core.fragment.ConstantLengthFragSet(3)

pyrosetta.rosetta.core.fragment.steal_constant_length_frag_set_from_pose(pose, confragset)

print(f'{filename}: {confragset.nr_frames()} frames')

fragio = pyrosetta.rosetta.core.fragment.FragmentIO()

fragio.write_data('***.frags', confragset)

The pose can have ligands (not sure about DNA), but true to `ignore_unrecognized_res` and false to `load_PDB_components` in the init would seem wise.

The output is something like the following, which I belive the same as normal fragment files.

FRAME    1   3

1     0 no_pdb BBTorsion S L      0.000   125.451   179.413

2     0 no_pdb BBTorsion S E   -158.982   169.226   179.818

3     0 no_pdb BBTorsion F E    -85.818   153.494   174.140

This works fine within PyRosetta. Although, one can use them without wasting time writing to disk, but often its via a vector of it, eg.

fragsets = pyrosetta.rosetta.utility.vector1_std_shared_ptr_core_fragment_FragSet_t(1)

fragsets[1] = confragset



## Example: Threading

align = pyrosetta.rosetta.core.sequence.read_aln(format='grishin', filename='***.grishin')

threader = pyrosetta.rosetta.protocols.comparative_modeling.ThreadingMover(align=align[1], template_pose=template_pose)

threader.build_loops(True)

threader.frag_libs(fragsets)

threader.apply(threaded_pose)

 

[Author]

Posts