I am hoping to 'fold' short phosphorylated peptide sequences, and started by generating fragments with SS predictions on the unphosphorylated peptides. For example, I ran this sequence through a local installation of PSIPRED and picked 200 3- and 9-mers:
>1
MKGDAHRYLAEFATG
I checked that the TYR_p:phosphorylated patch was active in the full-atom patches.txt, then created a centroid patch and added it to the patches.txt by analogy to existing patches:
NAME phosphorylated
TYPES PHOSPHORYLATION
BEGIN_SELECTOR
PROPERTY PROTEIN
AA TYR
NOT VARIANT_TYPE PHOSPHORYLATION ## Don't phosphorylate a residue twice
END_SELECTOR
BEGIN_CASE ### THE GENERAL CASE ##########################################
ADD_PROPERTY CHARGED ## For the phosphate group
END_CASE
If I modify the centroid patch, I get an error that a residue type cannot be found for TYR_p:phosphorylated, so it is definitely doing something.
I have then been attempting to run AbinitioRelax on the phosphorylated fasta:
>1
MKGDAHR[TYR_p:phosphorylated]LAEFATG
And get weird output and errors:
protocols.abinitio.AbrelaxApplication: read fasta sequence: 36 residues
MKGDAHR[TYR_p:phosphorylated]LAEFATG
...
core.chemical.Patch: Reading patch file: /export/ActiveUsers/kentgorday/rosetta_src_2018.33.60351_bundle/main/database/chemical/residue_type_sets/centroid/patches/tyr_phosphorylated.txt
...
core.pose: one_letter: MKGDAHRLAEFATG
core.pose: seq_in: MKGDAHR[TYR_p:phosphorylated]LAEFATG
core.chemical.ResidueTypeFinder: No ResidueTypes remain after filtering for matched variants.
core.chemical.ResidueTypeFinder: No ResidueTypes remain after filtering for matched variants.
...
core.pose: residue_types_from_sequence(): seqpos: 6 aa H HIS
core.chemical.PatchOperations.hh: AddProperty::apply: CHARGED
core.chemical.Patch: successfully patched: TYR to: TYR:phosphorylated
core.pose: residue_types_from_sequence(): seqpos: 7 aa R ARG
...
core.pose: make_pose_from_sequence(): seqpos: 6 HIS
core.conformation.Conformation: CURRENT_FOLD_TREE EDGE 1 6 -1
core.pose: make_pose_from_sequence(): seqpos: 7 TYR
core.conformation.Conformation: CURRENT_FOLD_TREE EDGE 1 7 -1
core.pose: make_pose_from_sequence(): seqpos: 8 LEU
core.conformation.Conformation: CURRENT_FOLD_TREE EDGE 1 8 -1
...
protocols.abinitio:
===================================================================
Stage 1
Folding with score0 for max of 20000
protocols.simple_moves.FragmentMover: 1 2 3 4 5 6 7size of insertmap: 7 -- insert_size: 9 9 9 9 9 9 9
protocols.simple_moves.FragmentMover: [ ERROR ] BEGIN: 7 SIZE: 9 TOTAL_RES: 14
protocols.simple_moves.FragmentMover: [ ERROR ] Are the fragments compatible with the fasta or the input PDB used to extract the folding sequence ?
protocols.simple_moves.FragmentMover: [ ERROR ] It appears that the fragments go up to residue 15 while the pose only has 14 residues!
Error: [ ERROR ] Must call shutdown() when finished using job distributor!
Which looks like somehow the arginine is being lost, even though the phosphotyrosine appears to be handled correctly.
Does anyone know what is causing this behavior, and whether this is an appropriate approach to ab intio for phosphopeptides (barring differences in secondary structure from Y to pY)?
Thank you for your help!
Kent
I think the issue is that you have the annotated sequence format slightly incorrect.
The part in the brackets are an addition to the one letter code, not the replacement. So instead of
MKGDAHR[TYR_p:phosphorylated]LAEFATG
which is attempting to make a 'R' residue with a full name of "TYR_p:phosphorylated", you would want
MKGDAHRY[TYR_p:phosphorylated]LAEFATG
(note the addition of the 'Y'.) This is why you're getting sequence size errors and confusion as to whether residue 7 is a tyrosine or an arginine.
Side note: the '_p' is no longer needed in specifying patch names. "TYR:phosphorylated" is the new way of writing it.