calc_CSP_RANK_for_directory_medoid.py

# calc_CSP_metrics.py
# calculate Fmeasure, Precision, Recall for each apo/holo pair with CSP data
# determine binding residues using UCBShift predictions for holo form, spectra aligned to holo shifts

import sys
from os.path import basename
from util import *
from paths import *
from os.path import exists, isfile, isdir, join
from os import listdir
from tqdm import tqdm

data_source_file = './CSPRANK.csv'
method = "MONTE"
z_value = 0

def align_shifts_to_seq(aligned_sequence, sequence, shifts):
    new_shifts = []
    seq_index = 0
    for i in range(len(aligned_sequence)):
        if seq_index < len(sequence) and aligned_sequence[i] == sequence[seq_index]:
            new_shifts.append(shifts[seq_index])
            seq_index += 1
        else:
            new_shifts.append(-1)
    return new_shifts

if __name__ == "__main__":
    if len(sys.argv) != 2:
        print("Usage: python calc_CSP_RANK_for_directory.py <bound>")
        sys.exit(1)

    bound = sys.argv[1].lower()

    directory = NMR_holo_structure_dir
    if isdir(directory) == False or len(listdir(directory)) == 0:
        print("Invalid directory")
        sys.exit(1)
    

    data = parse_csv(data_source_file)
    apos = [str(data['apo_bmrb']) for data in data]
    bounds = [data['holo_pdb'] for data in data]
    if bound not in bounds:
        print("Invalid bound")
        sys.exit(1)

    apo = apos[bounds.index(bound)]
    holo = bound
    pdb_files = [f for f in listdir(directory) if isfile(join(directory, f)) and f.endswith('.pdb') and f.find(holo) != -1]
    # medoid = find_medoid_structure(pdb_files)
    # medoid = medoid[medoid.rfind('/')+1:medoid.rfind('.')]
    # basenames = [f[f.rfind('/')+1:f.rfind('.')] for f in listdir(directory)]
    TP, FP, FN, TN = get_confusion(apo, holo, "MONTE", "UCBShift", "", "", structure_source = "NMR", basename=pdb_files)
    F, MCC, consensus = get_F_MCC_cons(TP, FP, FN, TN)

    print("F: ", F)
    print("MCC: ", MCC)
    print("consensus: ", consensus)