Upload 23 files

anti_mammalian_cells_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.1525790244340896
+1,0.0760533288121223
+2,0.1492621749639511
+3,0.1219991967082023
+4,0.1047019213438034
+5,0.1363092511892318
+6,0.0930725336074829
+7,0.0958817079663276
+8,0.1338892579078674
+9,0.0932814627885818

antibacterial_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.4496227502822876
+1,0.3883198201656341
+2,0.3562843203544616
+3,0.298247218132019
+4,0.4385020732879638
+5,0.4251146614551544
+6,0.4341979026794433
+7,0.2845156490802765
+8,0.3436572253704071
+9,0.3261722922325134

antibiofilm_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0010717363329604
+1,0.0004516514600254
+2,0.000373906281311
+3,0.0007363113109022
+4,3.484930857666768e-05
+5,0.0003151300188619
+6,0.0011047080624848
+7,0.0007902314537204
+8,0.0006953802076168
+9,0.0003028515202458

anticancer_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.1051242128014564
+1,0.0922926068305969
+2,0.0994595140218734
+3,0.0824442058801651
+4,0.0792957991361618
+5,0.072281502187252
+6,0.0954500809311866
+7,0.0436021983623504
+8,0.0734143629670143
+9,0.0542230121791362

anticandida_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0272127203643322
+1,0.0113312294706702
+2,0.0158187840133905
+3,0.009648754261434
+4,0.0055011077784001
+5,0.011053130030632
+6,0.0015367614105343
+7,0.0144472680985927
+8,0.0127825438976287
+9,0.01068951562047

antifungal_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.1905697137117386
+1,0.2031028866767883
+2,0.1090441793203353
+3,0.1683354675769806
+4,0.2166067212820053
+5,0.1459249407052993
+6,0.1769887655973434
+7,0.123556450009346
+8,0.1815296709537506
+9,0.1819620281457901

antigram-negative_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.350467324256897
+1,0.2841741442680359
+2,0.1829241067171096
+3,0.2051365971565246
+4,0.2392990440130233
+5,0.250076562166214
+6,0.2709463834762573
+7,0.3161788880825043
+8,0.2945723533630371
+9,0.3661049306392669

antigram-positive_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.159443125128746
+1,0.175675019621849
+2,0.1815598160028457
+3,0.2127318680286407
+4,0.1937155872583389
+5,0.1311574429273605
+6,0.2421468198299408
+7,0.164654865860939
+8,0.2011494487524032
+9,0.1703412532806396

antihiv_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.001297834329307
+1,0.0147092789411544
+2,0.0280767139047384
+3,0.0188311729580163
+4,0.0316063687205314
+5,0.0002793179592117
+6,0.0394531860947608
+7,0.0352290384471416
+8,0.0161530096083879
+9,0.0399153716862201

antimalarial_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0002518623368814
+1,0.0011312034912407
+2,0.0008880426757968
+3,0.000496806751471
+4,0.0012301608221605
+5,0.0008584852912463
+6,0.0002407601568847
+7,0.0007967063575051
+8,0.0017376942560076
+9,0.0015417954418808

antimrsa_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0141361905261874
+1,0.0036627338267862
+2,0.0077131749130785
+3,0.0193892307579517
+4,0.0269243717193603
+5,0.011976390145719
+6,0.014747904613614
+7,0.0004355635028332
+8,0.0149725144729018
+9,0.001316599198617

antiparasitic_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0133489668369293
+1,0.013213163241744
+2,0.0098364204168319
+3,0.0098508978262543
+4,0.0078889606520533
+5,0.0084595428779721
+6,0.0112555986270308
+7,0.009906081482768
+8,0.0203342288732528
+9,0.0079685244709253

antiplasmodial_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0001347773213637
+1,0.0009133622515946
+2,9.326116014563011e-06
+3,0.0001174091376014
+4,3.263696635258384e-05
+5,8.079609688138589e-05
+6,0.002609065035358
+7,0.0026866241823881
+8,0.0004498923080973
+9,4.718012496596202e-05

antiprotozoal_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.003057480789721
+1,0.0040516722947359
+2,0.0025351848453283
+3,0.0045190006494522
+4,0.0001927299890667
+5,0.000583435583394
+6,0.0027324347756803
+7,1.884532684925944e-05
+8,0.0009825642919167
+9,0.0006906135822646

antitb_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.000651930575259
+1,0.0006640456267632
+2,0.007218284998089
+3,0.0021461523137986
+4,0.0012260759249329
+5,0.0005806351546198
+6,0.0083108721300959
+7,0.0032738070003688
+8,0.0002077087265206
+9,0.0016663705464452

antiviral_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.2781128883361816
+1,0.2598320543766022
+2,0.2447118610143661
+3,0.2298509180545807
+4,0.3155396282672882
+5,0.3103104829788208
+6,0.2029070407152176
+7,0.2685596346855163
+8,0.2826785147190094
+9,0.2654517292976379

anurandefense_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.06325114518404
+1,0.040009070187807
+2,0.0438646599650383
+3,0.0655954405665397
+4,0.0161753576248884
+5,0.0274067930877208
+6,0.0275417882949113
+7,0.0296123120933771
+8,0.1663894206285476
+9,0.0152728063985705

app.py

@@ -0,0 +1,223 @@
+from Bio import SeqIO
+import os
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import torch.nn.functional as F
+from tqdm import tqdm
+import numpy as np
+import scipy.stats
+import pathlib
+import copy
+import time
+# from termcolor import colored
+import vocab
+from model import SequenceMultiTypeMultiCNN_1
+from tools import EarlyStopping
+from data_feature import Dataset
+from sklearn.metrics import roc_auc_score
+from sklearn.metrics import accuracy_score, f1_score, confusion_matrix, matthews_corrcoef
+import pandas as pd
+import argparse
+from tqdm import tqdm
+from io import StringIO
+import gradio as gr
+
+device = torch.device("cpu")
+
+
+def return_y(data_iter, net):
+    y_pred = []
+
+    all_seq = []
+    for batch in data_iter:
+        all_seq += batch['sequence']
+
+        AAI_feat = batch['seq_enc_AAI'].to(device)
+        onehot_feat = batch['seq_enc_onehot'].to(device)
+        BLOSUM62_feat = batch['seq_enc_BLOSUM62'].to(device)
+        PAAC_feat = batch['seq_enc_PAAC'].to(device)
+        # bert_feat=batch['seq_enc_bert'].to(device)
+        # bert_mask=batch['seq_enc_mask'].to(device)
+        outputs = net(AAI_feat, onehot_feat, BLOSUM62_feat, PAAC_feat)
+        # outputs = model(x)
+        y_pred.extend(outputs.cpu().numpy())
+
+    return y_pred, all_seq
+
+
+def testing(batch_size, patience, n_epochs, testfasta, seq_len, cdhit_value, cv_number, save_file, model_file):
+    model = SequenceMultiTypeMultiCNN_1(d_input=[531, 21, 23, 3], vocab_size=21, seq_len=seq_len,
+                                        dropout=0.1, d_another_h=128, k_cnn=[2, 3, 4, 5, 6], d_output=1).to(device)
+
+    dataset = Dataset(fasta=testfasta)
+    test_loader = dataset.get_dataloader(batch_size=batch_size, max_length=seq_len)
+
+    model.load_state_dict(torch.load(model_file, map_location=torch.device('cpu'))['state_dict'])
+    model.eval()
+    with torch.no_grad():
+        new_y_pred, all_seq = return_y(test_loader, model)
+
+    final_y_pred = copy.deepcopy(new_y_pred)
+
+    final_y_pred = np.array(final_y_pred).T[0].tolist()
+
+    pred_dict = {'seq': all_seq, 'predictions': final_y_pred}
+    pred_df = pd.DataFrame(pred_dict)
+    pred_df.to_csv(save_file, index=None)
+
+
+all_function_names = ['antibacterial', 'antigram-positive', 'antigram-negative', 'antifungal', 'antiviral', \
+                      'anti_mammalian_cells', 'antihiv', 'antibiofilm', 'anticancer', 'antimrsa', 'antiparasitic', \
+                      'hemolytic', 'chemotactic', 'antitb', 'anurandefense', 'cytotoxic', \
+                      'endotoxin', 'insecticidal', 'antimalarial', 'anticandida', 'antiplasmodial', 'antiprotozoal']
+
+
+# os.environ['CUDA_LAUNCH_BLOCKING'] = 1
+
+
+def predict(test_file):
+    # fas_id = []
+    fas_seq = [test_file]
+    # for seq_record in SeqIO.parse(test_file, "fasta"):
+    #     fas_seq.append(str(seq_record.seq).upper())
+    #     fas_id.append(str(seq_record.id))
+
+    seq_len = 200
+    batch_size = 32
+    cdhit_value = 40
+    vocab_size = len(vocab.AMINO_ACIDS)
+
+    epochs = 300
+    temp_save_AMP_filename = '%s ' % (time.strftime('%Y-%m-%d-%H-%M-%S', time.localtime()))
+    for cv_number in tqdm(range(10)):
+        testing(testfasta=fas_seq,
+                model_file=f'textcnn_cdhit_40_{cv_number}.pth.tar',
+                save_file=f'{temp_save_AMP_filename}_{cv_number}.csv',
+                batch_size=batch_size, patience=10, n_epochs=epochs, seq_len=seq_len, cdhit_value=cdhit_value
+                , cv_number=cv_number)
+
+    pred_prob = []
+    for cv_number in tqdm(range(10)):
+        df = pd.read_csv(f'{temp_save_AMP_filename}_{cv_number}.csv')
+        data = df.values.tolist()
+        temp = []
+        for i in tqdm(range(len(data))):
+            temp.append(data[i][1])
+        pred_prob.append(temp)
+    pred_prob = np.average(pred_prob, 0)
+    pred_AMP_label = []
+    for i in tqdm(range(len(pred_prob))):
+        if pred_prob[i] > 0.5:
+            pred_AMP_label.append('Yes')
+        else:
+            pred_AMP_label.append('No')
+
+    for function_name in all_function_names:
+        temp_dir_list = os.listdir('tmp_save')
+        if function_name not in temp_dir_list:
+            os.mkdir( function_name)
+        for cv_number in tqdm(range(10)):
+            testing(testfasta=fas_seq,
+                    model_file=f'{function_name}textcnn_cdhit_100_0.pth.tar',
+                    save_file=f'{function_name}{temp_save_AMP_filename}_{cv_number}.csv',
+                    batch_size=batch_size, patience=10, n_epochs=epochs, seq_len=seq_len, cdhit_value=cdhit_value
+                    , cv_number=cv_number)
+
+    all_function_pred_label = []
+    for function_name in all_function_names:
+
+        function_threshold_df = pd.read_csv(f'{function_name}_yd_threshold.csv', index_col=0)
+        function_thresholds = function_threshold_df.values[:, 0]
+
+        each_function_data = []
+
+        for cv_number in tqdm(range(10)):
+            df = pd.read_csv(f'{function_name}{temp_save_AMP_filename}_{cv_number}.csv')
+            data = df.values.tolist()
+            temp = []
+            for i in tqdm(range(len(data))):
+
+                if data[i][1] > function_thresholds[cv_number]:
+                    temp.append(1)
+                else:
+                    temp.append(0)
+            each_function_data.append(temp)
+        each_function_data = np.average(each_function_data, 0)
+        pred_each_function_label = []
+        for i in tqdm(range(len(each_function_data))):
+            if each_function_data[i] > 0.5:
+                pred_each_function_label.append('Yes')
+            else:
+                pred_each_function_label.append('No')
+
+        all_function_pred_label.append(pred_each_function_label)
+
+    all_function_cols = ['antibacterial', 'anti-Gram-positive', 'anti-Gram-negative', 'antifungal', 'antiviral', \
+                         'anti-mammalian-cells', 'anti-HIV', 'antibiofilm', 'anticancer', 'anti-MRSA', 'antiparasitic', \
+                         'hemolytic', 'chemotactic', 'anti-TB', 'anurandefense', 'cytotoxic', \
+                         'endotoxin', 'insecticidal', 'antimalarial', 'anticandida', 'antiplasmodial', 'antiprotozoal']
+
+    pred_contents_dict = {'sequence': fas_seq, 'AMP': pred_AMP_label}
+    for i in tqdm(range(len(all_function_cols))):
+        pred_contents_dict[all_function_cols[i]] = all_function_pred_label[i]
+
+    pred_contents_df = pd.DataFrame(pred_contents_dict)
+
+    for function_name in all_function_names:
+        for cv_number in tqdm(range(10)):
+            os.remove(f'{function_name}{temp_save_AMP_filename}_{cv_number}.csv')
+    for cv_number in tqdm(range(10)):
+        os.remove(f'{temp_save_AMP_filename}_{cv_number}.csv')
+    result_csv = pd.DataFrame({ 'Prediction': pred_AMP_label})
+    result_csv_string = StringIO()
+    result_csv.to_csv(result_csv_string, index=False)
+    result_csv_string.seek(0)
+
+    return pred_contents_df
+    # master.insert_one({'Test Report': res_val})
+
+
+if __name__ == '__main__':
+    pd.set_option('display.max_columns', None)
+    pd.set_option('display.max_colwidth', -1)
+
+    # parser = argparse.ArgumentParser(description='proposed model')
+
+    # parser.add_argument('-output_file_name', default='prediction_output', type=str)
+
+    # parser.add_argument('-test_fasta_file', default='examples/samples.fasta', type=str)
+    # args = parser.parse_args()
+
+    # output_file_name = args.output_file_name
+    # test_file = args.test_fasta_file
+    # flag = 0
+    # for seq_record in SeqIO.parse(test_file, "fasta"):
+    #     temp_id = str(seq_record.id)
+    #     temp_seq = str(seq_record.seq)
+    #     if len(set(temp_seq.upper()).difference(set('ACDEFGHIKLMNPQRSTVWY'))) > 0:
+    #         flag = 1
+    #         print('input error: have unusual amino acids')
+    #         break
+
+    # if flag == 0:
+    # pred_df = predict(test_file)
+    # pred_df.to_csv(output_file_name + '.csv')
+    with gr.Blocks() as demo:
+        gr.Markdown(
+            """
+            
+        # Welcome to Antimicrobial Peptide Attribute Prediction Model
+
+        This is an online model for predicting attributes of antimicrobial peptides. Here, you can simply input a protein sequence, such as QGLFFLGAKLFYLLTLFL, and the model will generate predictions for various attributes.
+
+        Please note that due to server limitations, large-scale predictions may not be supported online. If you have a need for large-scale predictions, I can provide you with the code or assist you with the predictions directly, free of charge. Feel free to contact me for any inquiries:
+        
+        Email: [email protected]
+
+        Let's get started!
+        
+            """)
+
+        iface = gr.Interface(fn=predict, inputs="text", outputs="text")
+    demo.launch(server_name='127.0.0.1', server_port=7788)

chemotactic_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0060868617147207
+1,9.572660201229156e-05
+2,0.0008185741025954
+3,9.364341531181708e-05
+4,0.0002393810573266
+5,0.0022423025220632
+6,0.0084763616323471
+7,0.0027119170408695
+8,3.411575744394213e-05
+9,0.0029187819454818

cytotoxic_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0080689117312431
+1,0.0061783450655639
+2,0.0107948025688529
+3,0.012933705933392
+4,0.0074981972575187
+5,0.000167274614796
+6,0.0103498054668307
+7,0.0073629915714263
+8,0.0012240845244377
+9,0.0001633148203836

endotoxin_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0015385682927444
+1,0.0006175263551995
+2,0.0014237745199352
+3,0.0001627063029445
+4,0.001902371761389
+5,0.0006754832575097
+6,0.0008196207927539
+7,0.0021725625265389
+8,0.0004220827540848
+9,0.0006830523489043

hemolytic_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0221392400562763
+1,0.0695223584771156
+2,0.0811304897069931
+3,0.0350562147796154
+4,0.0791068449616432
+5,0.0475858710706234
+6,0.0865858867764473
+7,0.0352442301809787
+8,0.0357107110321521
+9,0.0703328251838684

insecticidal_yd_threshold.csv

@@ -0,0 +1,11 @@
+,threshold
+0,0.0042725815437734
+1,0.0039733867160975
+2,0.0020466167479753
+3,0.0001893758308142
+4,0.005238143261522
+5,0.003733716905117
+6,0.0044953846372663
+7,0.0027529671788215
+8,0.0010488195111975
+9,0.0006791127379983