Chou Fasman Protein Secondary Structure Prediction - Python Implementation

This code implements the Chou Fasman method to predict the secondary structure (helices and strands) of a protein sequence.

Explanation (only for code.py):

1. Data and Initialization:

   • A dictionary amino_acid_values stores scores for each amino acid related to helix (P(h)) and strand (P(s)) propensity.
   • A protein sequence sequence_given is defined.
   • Three lists are initialized:
     • List_Containing_Possible_Helices: Stores predicted helix residues ('-'' for non-helix).
     • List_Containing_Possible_Betas: Stores predicted strand residues ('-'' for non-strand).
     • output: Stores the final predicted secondary structure ('-'' for undetermined).

2. Scoring Functions:

   • Helix_Score_using_P(residue): Calculates the total helix propensity score for a given amino acid sequence residue.
   • Beta_Score_using_P(residue): Calculates the total strand propensity score for a given amino acid sequence residue.

3. Nucleation Site Calculation:

   • calculate(i, type): Analyzes a window of amino acids around position i to determine if it's a valid nucleation site for a helix or strand based on the type ("S" for strand, "H" for helix).

4. Extension Functions:

   • extend_left(j, type): Extends a predicted helix/strand region (type) to the left from position j as long as the P(h)/P(s) score remains favorable.
   • extend_right(j, type): Extends a predicted helix/strand region (type) to the right from position j as long as the P(h)/P(s) score remains favorable.

5. Sequence Processing:

   • process_sequence(sequence, target, length): Processes the entire sequence recursively. It calls calculate to identify potential nucleation sites for the target secondary structure (target, "H" for helix, "S" for strand) with a minimum length of length. If a valid site is found, it extends the predicted region using extend_left and extend_right.

6. Conflict Resolution:

   • A dictionary dict_conflicts_resolve is created to store predicted helix and strand characters at each position.
   • A list Conflicting_Regions stores the start and end indices of regions with conflicting helix and strand predictions.
   • A list resolved_List is created to store the final predicted secondary structure for each amino acid.
   • The code iterates through the sequence and resolves conflicts based on the following logic:
     • If neither helix nor strand is predicted, the residue is marked as '-'.
     • If only one (helix or strand) is predicted, the residue is marked accordingly.
     • If both helix and strand are predicted for a region, the amino acid sequence within that region is evaluated using the scoring functions Helix_Score_using_P and Beta_Score_using_P. The structure with the higher score is assigned to the conflicting region in resolved_List. If the scores are equal, the code prioritizes strands by default.

7. Output:

   • The code prints the following information:
     • Predicted helices only.
     • Predicted strands only.
     • List of conflicting regions.
     • Final predicted secondary structure after resolving conflicts.

Overall, this code implements the Chou Fasman method to predict the secondary structure of a protein sequence. It considers the amino acid propensities for helix and strand formation and employs a window-based approach to identify potential secondary structure elements.

Note: The Chou Fasman method is a relatively simple approach with limitations. More sophisticated methods might offer higher accuracy for protein secondary structure prediction.