How can I use ss2 or FASTA files as inputs for protein-folding simulation using the MARTINI course-grained model?

Question

I asked this at Biostars but the post seems to have been deleted there.

My professor told me to implement MARTINI ForceField-based coarse-grained model to simulate 2GB1A protein-folding using Python-3.5 and to use *.ss2 and *.fasta_ss files as inputs.

Thus far, I have used manual coordinates for protein simulation. I never even used PDB files as inputs. So, I have no idea how the mentioned file formats can be used for simulation.

Do I need to extract point coordinates from *.ss2 and *.fasta_ss files? If so, how?

---

This is the link of PSRIRED VFORMAT.

The following is the content of *.ss2 file of 2GB1A protein:

# PSIPRED VFORMAT

   1 M C   1.000  0.000  0.000
   2 T E   0.000  0.000  1.000
   3 Y E   0.000  0.000  1.000
   4 K E   0.000  0.000  1.000
   5 L E   0.000  0.000  1.000
   6 I E   0.000  0.000  1.000
   7 L E   0.000  0.000  1.000
   8 N C   1.000  0.000  0.000
   9 G C   1.000  0.000  0.000
  10 K C   1.000  0.000  0.000
  11 T C   1.000  0.000  0.000
  12 L C   1.000  0.000  0.000
  13 K C   1.000  0.000  0.000
  14 G E   0.000  0.000  1.000
  15 E E   0.000  0.000  1.000
  16 T E   0.000  0.000  1.000
  17 T E   0.000  0.000  1.000
  18 T E   0.000  0.000  1.000
  19 E E   0.000  0.000  1.000
  20 A C   1.000  0.000  0.000
  21 V C   1.000  0.000  0.000
  22 D C   1.000  0.000  0.000
  23 A H   0.000  1.000  0.000
  24 A H   0.000  1.000  0.000
  25 T H   0.000  1.000  0.000
  26 A H   0.000  1.000  0.000
  27 E H   0.000  1.000  0.000
  28 K H   0.000  1.000  0.000
  29 V H   0.000  1.000  0.000
  30 F H   0.000  1.000  0.000
  31 K H   0.000  1.000  0.000
  32 Q H   0.000  1.000  0.000
  33 Y H   0.000  1.000  0.000
  34 A H   0.000  1.000  0.000
  35 N H   0.000  1.000  0.000
  36 D C   1.000  0.000  0.000
  37 N C   1.000  0.000  0.000
  38 G C   1.000  0.000  0.000
  39 V C   1.000  0.000  0.000
  40 D C   1.000  0.000  0.000
  41 G C   1.000  0.000  0.000
  42 E E   0.000  0.000  1.000
  43 W E   0.000  0.000  1.000
  44 T E   0.000  0.000  1.000
  45 Y E   0.000  0.000  1.000
  46 D E   0.000  0.000  1.000
  47 D C   1.000  0.000  0.000
  48 A C   1.000  0.000  0.000
  49 T C   1.000  0.000  0.000
  50 K C   1.000  0.000  0.000
  51 T E   0.000  0.000  1.000
  52 F E   0.000  0.000  1.000
  53 T E   0.000  0.000  1.000
  54 V E   0.000  0.000  1.000
  55 T E   0.000  0.000  1.000
  56 E C   1.000  0.000  0.000

Its corresponding fasta file is as follows (I haven't found a fasta_ss example):

>2gb1A
MTYKLILNGKTLKGETTTEAVDAATAEKVFKQYANDNGVDGEWTYDDATKTFTVTE

Answer 1

In order to run Molecular Dynamics simulations (independent of the force-field and type of simulation) you need a file with the structure of your protein: type of atoms, residues and (X, Y, Z) coordinates.

Normally, the starting file with all this information is in PDB format like this example.

You can search the RCSB Protein Data Base and, if you are lucky, found the structure of your protein. There you will get structures determined by X-ray crystallography, NMR and CryoEM. If you aren't lucky, then you need to use the sequence (in FASTA format, for example) and generate its structure using homology modeling. Right now, the RCSB Protein Data Base also has available Computed Structure Models generated by AlphaFoldDB and RoseTTAFold.

Once you have the PDB file, you can go to online services like CHARMM-GUI and generate the inputs for running Molecular Dynamic simulations for both all-atom and coarse-grained molecular dynamic.