595693085 / DGraphDTA

a novel DTA predition method using graph neural network
70 stars 40 forks source link

PSSM issue (see PR #15) #16

Open jyaacoub opened 1 year ago

jyaacoub commented 1 year ago

More details in #15

The issue is in data_process.py line 149 (PSSM_calculation()):

def PSSM_calculation(aln_file, pro_seq):
    pfm_mat = np.zeros((len(pro_res_table), len(pro_seq)))
    with open(aln_file, 'r') as f:
        line_count = len(f.readlines()) # CANT GET THE LINE COUNT LIKE THIS BC THE FILE OBJECT BECOMES EMPTY AFTER THE FIRST READ
        for line in f.readlines(): # ERROR: BIG ISSUE HERE - f.readlines() is empty. FOR LOOP NEVER EXECUTES
            if len(line) != len(pro_seq):
                print('error', len(line), len(pro_seq))
                continue
            count = 0
595693085 commented 1 year ago

Thanks for your suggestion. Maybe the experiment be influenced by accidental factors. We will check and correct it. and in our current work, the PSSM feature is removed. Thank you again for your interest in our work, and you are truly a rigorous and serious researcher!