Abinitio starts at the 27th amino acid

[Anonymous]

Hi all,

I am trying to predict 2XGF, however, it seems Rosetta, for some reason, did not start from the first.

---

~/2XGF$ ~/rosetta3.4/rosetta_source/bin/AbinitioRelax.linuxgccrelease -in:file:native ~/2XGF/2XGF.pdb -in:file:frag3 ~/2XGF/aat000_03_05.200_v1_3 -in:file:frag9 ~/2XGF/aat000_09_05.200_v1_3 -database ~/rosetta3.4/rosetta_database -abinitio:relax -relax:fast -use_filters true -psipred_ss2 ~/2XGF/t000_.psipred_ss2 -kill_hairpins ~/2XGF/t000_.psipred_ss2 -nstruct 2000 -out:file:silent 2XGF_silent.out -out:path ~/2XGF/ -ignore_unrecognized_res
core.init: Mini-Rosetta version unknown from unknown
core.init: command: /home/k1064285/rosetta3.4/rosetta_source/bin/AbinitioRelax.linuxgccrelease -in:file:native /home/k1064285/2XGF/2XGF.pdb -in:file:fasta /home/k1064285/2XGF/2XGF.fasta.txt -in:file:frag3 /home/k1064285/2XGF/aat000_03_05.200_v1_3 -in:file:frag9 /home/k1064285/2XGF/aat000_09_05.200_v1_3 -database /home/k1064285/rosetta3.4/rosetta_database -abinitio:relax -relax:fast -use_filters true -psipred_ss2 /home/k1064285/2XGF/t000_.psipred_ss2 -kill_hairpins /home/k1064285/2XGF/t000_.psipred_ss2 -nstruct 2000 -out:file:silent 2XGF_silent.out -out:path /home/k1064285/2XGF/ -ignore_unrecognized_res
core.init: ‘RNG device’ seed mode, using ‘/dev/urandom’, seed=-1941539611 seed_offset=0 real_seed=-1941539611
core.init.random: RandomGenerator:init: Normal mode, seed=-1941539611 RG_type=mt19937
core.chemical.ResidueTypeSet: Finished initializing fa_standard residue type set. Created 6225 residue types
core.pack.task: Packer task: initialize from command line()
core.chemical.ResidueTypeSet: Finished initializing centroid residue type set. Created 1980 residue types
SSYPIGAPIPWPSDSVPAGFALMEGQTFDKSAYPKLAVAYPSGVIPDMRGQTIKGKPSGRAVLSAEADGVKAHSHSASASSTDLGTKTTSSFDYGTKGTNSTGGHTHSGSGSTSTNGEHSHYIEAWNGTGVGGNKMSSYAISYRAGGSNTNAAGNHSHTFSFGTSSAGDHSHSVGIGAHTHTVAIGSHGHTITVNSTGNTENTVKNIAFNYIVRLAZZZZZZZSSYPIGAPIPWPSDSVPAGFALMEGQTFDKSAYPKLAVAYPSGVIPDMRGQTIKGKPSGRAVLSAEADGVKAHSHSASASSTDLGTKTTSSFDYGTKGTNSTGGHTHSGSGSTSTNGEHSHYIEAWNGTGVGGNKMSSYAISYRAGGSNTNAAGNHSHTFSFGTSSAGDHSHSVGIGAHTHTVAIGSHGHTITVNSTGNTENTVKNIAFNYIVRLASSYPIGAPIPWPSDSVPAGFALMEGQTFDKSAYPKLAVAYPSGVIPDMRGQTIKGKPSGRAVLSAEADGVKAHSHSASASSTDLGTKTTSSFDYGTKGTNSTGGHTHSGSGSTSTNGEHSHYIEAWNGTGVGGNKMSSYAISYRAGGSNTNAAGNHSHTFSFGTSSAGDHSHSVGIGAHTHTVAIGSHGHTITVNSTGNTENTVKNIAFNYIVRLA
can not find a residue type that matches the residue FE2at position 217

ERROR: core::util::switch_to_residue_type_set fails

ERROR:: Exit from: src/core/util/SwitchResidueTypeSet.cc line: 143

---

the sequence appears to start at SSYPI…, however, S should be the 27th amino acid not the first.

Any help?

Jad

[Anonymous]

I’m guessing the sequence related issue is due to the fact that the 2XGF structure as obtained from the PDB starts at SSYPI. If I recall correctly, the application assumes that the native structure provided will have the same size and length as the sequence you’re modeling. You may need to “fake” the first 26 amino acids.

The error that’s halting the execution, though, is that there isn’t an iron parameters file in centroid mode. The iron atoms in the PDB file can get read in as a full-atom mode, but then when the residues get transformed into centroid mode, it doesn’t work. I’d recommend using the clean_PDB.py cleanup script to clean up the native PDB and remove the iron atoms from the PDB (or at least do it manually.)

[Anonymous]

Rmoretti is right, Rosetta assumes that the sequence and reference PDB have an identical sequence, the numbering is not important.
There are quite a few PDB structures where the Fasta sequence is longer than the sequence in the PDB file. What I learned is that the safest way is to extract the Fasta sequence from the PDB file. If you already created the decoys I found two possibilities to calculate the RMSD (using python scripts), either cut the excessive amino acids off and compare the truncated decoys or use a a program like pymol which aligns the structures first and then calculate the RMSD.

[Anonymous]

Thank you a lot.

[Author]

Posts