Initial commit

.gitattributes

+SSA_embed.model filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text

README.md

+## prPred-DRLF
+prPred-DRLF is a tool to identify the plant resistance proteins (R proteins) based on deep representation learning features
+
+prPred-DRLF is an open-source Python-based toolkit, which operates depending on the Python environment (##Python Version 3.7). 
+
+### **If your computer has GPU,it will be faster.  **
+
+
+### **Download**
+
+    git clone git@github.com:Wangys-prog/prPred-DRLF.git
+    
+
+### **Install dependencies**
+
+    pip3 install -r requirements.txt 
+    or 
+    pip3 install joblib==1.0.1  
+    pip3 install tape_proteins==0.5
+    pip3 install numpy==1.19.2 
+    pip3 install pandas==1.2.0 
+    pip3 install Bio==0.4.1 
+    pip3 install lightgbm-2.3.0
+    
+#### For python3.7 
+#### If you have GPU  # CUDA 9.2   
+    pip install torch==1.2.0 torchvision==0.4.0 
+#### If you have GPU CUDA 10.0  
+    pip install torch==1.2.0+cu92 torchvision==0.4.0+cu92 -f https://download.pytorch.org/whl/torch_stable.html
+#### CPU Only python3.7
+    pip install torch==1.2.0+cpu torchvision==0.4.0+cpu -f https://download.pytorch.org/whl/torch_stable.html 
+
+### Input parameters
+
+    python3 prPred-DRLF.py -h  
+    $ usage: Script for predicting plant R protein using deep representation learning features  
+    $       [-h] [-i I] [-o O]  
+    $ optional arguments:  
+    $  -h, --help  show this help message and exit  
+    $  -i I        input sequences in Fasta format  
+    $  -o O        path to saved CSV file  
+    
+    python3 prPred-DRLF.py -i ./dataset/test_data.fasta -o ./dataset/predict_result
+
+
+### Webserver
+
+    http://lab.malab.cn/soft/prPred-DRLF
+
+### Other tools
+
+    http://lab.malab.cn/~wys/
+
+### Please cite
+    (1) Yansu Wang, Lei Xu, Quan Zou, Chen Lin. prPred-DRLF: plant R protein predictor using deep representation learning features. Proteomics. 2021. DOI: 10.1002/pmic.202100161
+    (2) Yansu Wang, Murong Zhou, Quan Zou, Lei Xu. Machine learning for phytopathology: from the molecular scale towards the network scale. Briefings in Bioinformatics. 2021, Doi:10.1093/bib/bbab037
+    (3) Wang Y, Wang P, Guo Y, et al. prPred: A Predictor to Identify Plant Resistance Proteins by Incorporating k-Spaced Amino Acid (Group) Pairs. Frontiers in bioengineering and biotechnology, 2021, 8: 1593.

dataset/predict_result

+,Sequence_ID,R protein possibility
+0,negative212|0,0.002
+1,positive29|1,1.0

dataset/test_data.fasta

+>negative212|0
+MGSQAGREASIFDSAVKGDTTILDRANIAKYGESSYLSHTPSLGNNIIHIAARRGKTTFVDAAIRLFPDLLWQKNNNDNTVVHEAAAEMGTKDCVRLLISYLKSSMSSSTTSSVSDHNRSNATAIAVPFLLERNCDGETPFHVALRCRNLAAAEELFADVKTHQQILLIKNNSGETPLHLYARYCAGGTFPKDLEHDDNHNVVDKPKFIDALIISNSAAIFEQDSDGFIPVMRAAQYGRVFAVLRMLLSYKQSTECRDLKGMNVLHHLRLRIADLNARDKDYTFWICEKILELHGVDTLIFLQDKDGNTPLHLAIMDRDYDIAELILKRYVEAYLKERKELIVSIKNKEGKTVLDLITFVPDIPITLRKLMEQSSMGCMMPEELYEAAKMGYVGVFGPQTSEVSNLSQETGSTTKNTQDDEYFLSQDVDGRNILHIALEHEREVFITTAIKSFPNLMYHKDSNGDTPLHIAARLRSNHSAYELLRESFHYWYANNRQSYDVTLRVPPWKVTNSRSNTPLHEAARTSNYSSLYTFIANCITPEIRSEAMSDVNEDGETVLHLIARYGQKLDAVLDHEHQHSLSVTEVLRESITSVYMRDRDGLTPILRAAHCGQIWVVARIGKMYPKSIYISDYKGKTVLHHLCSSALDVVNELEIVLQLWEQVSETFPNGDDLMFSQDHDGNTPLHLAIMAQNFRVVHYFIRYYIKVKSSLRNELMGLMNNDGKSIMNLITCGSDIPPNLKKLIKQASMETPMEDDIYEAATKGDINVINEHNRLQCQAYDGSNILHIAVRHRKLSLLMAVVEKNYLYHLIYWEDSKGDTPLHVAAGIQAPEALQFVEQCVKKWKPFYGLLPWTVRNMKGNTVLHEAARFCNYEVILSALKYATDINDVKLNQKIIKDVNDDGETVLHIIARYATYKAKDFLEKVSRELKSLVYMRDYEGFTPVLRAVQCGRLGVARLLVQRYPQSVEIADNKGRTILHHMRSLVVDLVDDLKDIVPVWKDILNYPEADNLRNAQDEDGNTPLHLSIADANVAKAKFLIERCLESTNKQELDINNNDGHTVYDLLSSRSHIPATKELQQLVSRSYVKERALVYDMMEDDLYNAAVKGDVQIFVKVPMDLESQPPTSVEAYFCRQTPGGSNIIHIALRHGSPKVKQFVKTALTHYPILSMRPDNNGDTPLHLAAKWKTGLSSVEVLIEASKCFINELGESDKAFYVAPWKVKNYKGNFPIHEALQSNNLKAAENLLGCDVEATSRVNDLGETPLHAFAKNGFAINNKKEAEKFVEKLIIAKTKEDESSNTASYIRDDEGLNQNAYNSAYIQDDEGLSQNAYNSAYIRDDEGLTPLLRAARSGRLEVVRAILTHCPQSAYLRDPWGRTFLHLLRFTGEDIDESLDGNFQKTGKELFVLREADSQRLVQDYEGNTPLHYAIKTQNSIAAIVLTQRCLEDEEHKELGLVNRDGQTVLDLLALHDVPSEIIQQIRKKLPKEVYLARSSYGIRNTETNQSANALSVVAALLATITFAAGLQVPGGFDSDDGSPVLLKNAVFAAFMVANTIAMCCSMLCLFLLLWVGIRKSHGSLMILDISIILLEFSFYLTMLAFTLGVFVVTLQKSLWLAVLVCVLSFITFLLTWKCSIKLVIKFVESVVAAGKKLASCCASKSADTEGACVGDNRTEK
+>positive29|1
+MAEILLTSVINKSVEIAGNLLIQEGKRLYWLKEDIDWLQREMRHIRSYVDNAKAKEAGGDSRVKNLLKDIQELAGDVEDLLDDFLPKIQQSNKFNYCLKRSSFADEFAMEIEKIKRRVVDIDRIRKTYNIIDTDNNNDDCVLLDRRRLFLHADETEIIGLDDDFNMLQAKLLNQDLHYGVVSIVGMPGLGKTTLAKKLYRLIRDQFECSGLVYVSQQPRASEILLDIAKQIGLTEQKMKENLEDNLRSLLKIKRYVFLLDDIWDVEIWDDLKLVLPECDSKVGSRIIITSRNSNVGRYIGGESSLHALQPLESEKSFELFTKKIFNFDDNNSWANASPDLVNIGRNIVGRCGGIPLAIVVTAGMLRARERTEHAWNRVLESMGHKVQDGCAKVLALSYNDLPIASRPCFLYFGLYPEDHEIRAFDLINMWIAEKFIVVNSGNRREAEDLAEDVLNDLVSRNLIQLAKRTYNGRISSCRIHDLLHSLCVDLAKESNFFHTAHDAFGDPGNVARLRRITFYSDNVMIEFFRSNPKLEKLRVLFCFAKDPSIFSHMAYFDFKLLHTLVVVMSQSFQAYVTIPSKFGNMTCLRYLRLEGNICGKLPNSIVKLTRLETIDIDRRSLIQPPSGVWESKHLRHLCYRDYGQACNSCFSISSFYPNIYSLHPNNLQTLMWIPDKFFEPRLLHRLINLRKLGILGVSNSTVKMLSIFSPVLKALEVLKLSFSSDPSEQIKLSSYPHIAKLHLNVNRTMALNSQSFPPNLIKLTLANFTVDRYILAVLKTFPKLRKLKMFICKYNEEKMDLSGEANGYSFPQLEVLHIHSPNGLSEVTCTDDVSMPKLKKLLLTGFHCRISLSERLKKLSK

prPred-DRLF.py

+import joblib
+import pandas as pd
+from prediction.get_feature import fasta,GnerateFeatures
+import argparse
+
+def predict(inputfasta,outfile):
+    seq_dict,id_list,seq_list = fasta(inputfasta)
+    feature_sel= GnerateFeatures(seq_dict)
+    model = joblib.load(open('./prediction/BiLSTM_unirep_model_lgb.pkl', 'rb'))
+    x=pd.read_csv("./prediction/BiLSTM_unirep_lgbm_feature.csv")
+    x2=x.iloc[:,1:]
+    y=pd.read_csv("./prediction/train_label.csv")
+    y2=y.iloc[:,1:]
+    model.fit(x2,y2)
+    pred_proba = model.predict_proba(feature_sel)
+    id3=[]
+    for i in range(len(id_list)):
+        id2 = []
+        id2.append(str(id_list[i]))
+        id2.append(str(round(float(pred_proba[:,1][i]),3)))
+        id3.append(id2)
+    col = ["Sequence_ID", "R protein possibility"]
+    result2=pd.DataFrame(data=id3,columns=col)
+    result2.to_csv(outfile)
+
+if __name__=="__main__":
+    parser = argparse.ArgumentParser(
+        'Script for predicting plant R protein using deep representation learning features')
+
+    parser.add_argument('-i', type=str, help='input sequences in Fasta format')
+    parser.add_argument('-o', type=str, help='path to saved CSV file')
+
+    args = parser.parse_args()
+    inputfasta= args.i
+    outfile = args.o
+    predict(inputfasta,outfile)
+

embbed_models @ 042c4d78

+Subproject commit 042c4d787140a2504b74b1cdb19c5586ba1c170f

prediction/feature_id.csv

+UniRep_F232,BiLSTM_F3253,BiLSTM_F231,UniRep_F1029,UniRep_F1338,UniRep_F747,BiLSTM_F750,BiLSTM_F2662,UniRep_F774,UniRep_F1757,UniRep_F650,UniRep_F1574,BiLSTM_F1420,BiLSTM_F77,UniRep_F662,BiLSTM_F1270,BiLSTM_F1404,BiLSTM_F1775,BiLSTM_F213,UniRep_F57,BiLSTM_F101,UniRep_F1800,UniRep_F1827,UniRep_F682,BiLSTM_F1758,UniRep_F204,UniRep_F861,BiLSTM_F2557,UniRep_F1793,BiLSTM_F655,BiLSTM_F1786,BiLSTM_F853,BiLSTM_F1448,UniRep_F4,BiLSTM_F671,UniRep_F510,BiLSTM_F1116,UniRep_F1754,BiLSTM_F482,UniRep_F212,BiLSTM_F1526,UniRep_F945,UniRep_F1573,BiLSTM_F1912,UniRep_F1241,BiLSTM_F1677,BiLSTM_F74,BiLSTM_F1696,BiLSTM_F524,UniRep_F814,BiLSTM_F1124,BiLSTM_F528,UniRep_F1299,UniRep_F408,UniRep_F1736,BiLSTM_F415,UniRep_F868,BiLSTM_F596,UniRep_F1818,UniRep_F1369,UniRep_F529,UniRep_F472,UniRep_F165,UniRep_F860,UniRep_F1376,BiLSTM_F3362,UniRep_F1214,BiLSTM_F22,UniRep_F1187,UniRep_F1182,UniRep_F1576,UniRep_F156,UniRep_F1095,UniRep_F1082,BiLSTM_F2843,UniRep_F1780,UniRep_F322,UniRep_F1791,BiLSTM_F154,BiLSTM_F252,UniRep_F51,BiLSTM_F3119,UniRep_F1843,UniRep_F660,UniRep_F1459,BiLSTM_F3485,BiLSTM_F2284,BiLSTM_F974,UniRep_F1336,BiLSTM_F1002,BiLSTM_F3578,BiLSTM_F865,BiLSTM_F969,BiLSTM_F3463,BiLSTM_F1066,UniRep_F657,BiLSTM_F1233,BiLSTM_F2256,UniRep_F538,UniRep_F1389,BiLSTM_F1481,BiLSTM_F1997,UniRep_F72,UniRep_F314,UniRep_F288,UniRep_F1625,UniRep_F97,BiLSTM_F1631,UniRep_F237,BiLSTM_F1705,BiLSTM_F1707,BiLSTM_F1740,BiLSTM_F1755,UniRep_F201,UniRep_F1423,UniRep_F183,BiLSTM_F2003,UniRep_F337,BiLSTM_F1424,UniRep_F49,BiLSTM_F1096,BiLSTM_F2179,BiLSTM_F1186,UniRep_F1711,UniRep_F1709,UniRep_F25,UniRep_F32,UniRep_F1581,UniRep_F1615,UniRep_F409,BiLSTM_F2134,BiLSTM_F2117,BiLSTM_F1339,UniRep_F362,BiLSTM_F2097,BiLSTM_F1414,UniRep_F1335,UniRep_F1229,UniRep_F1024,BiLSTM_F75,UniRep_F980,UniRep_F1877,BiLSTM_F2869,UniRep_F1840,BiLSTM_F87,BiLSTM_F525,BiLSTM_F2725,BiLSTM_F336,UniRep_F1065,BiLSTM_F2711,UniRep_F879,UniRep_F1817,UniRep_F889,BiLSTM_F3019,UniRep_F1090,BiLSTM_F3181,UniRep_F902,UniRep_F904,BiLSTM_F370,UniRep_F1832,UniRep_F1093,UniRep_F925,BiLSTM_F110,BiLSTM_F2747,BiLSTM_F2877,BiLSTM_F273,BiLSTM_F2879,BiLSTM_F166,BiLSTM_F794,UniRep_F703,UniRep_F704,BiLSTM_F2,BiLSTM_F3382,UniRep_F749,UniRep_F770,UniRep_F1215,BiLSTM_F2837,BiLSTM_F38,UniRep_F919,UniRep_F790,UniRep_F1539,BiLSTM_F3312,BiLSTM_F696,BiLSTM_F2839,BiLSTM_F677,UniRep_F1478,UniRep_F1149,BiLSTM_F53,BiLSTM_F3114,BiLSTM_F3051,BiLSTM_F3053,UniRep_F62,UniRep_F73,BiLSTM_F3423,UniRep_F92,UniRep_F129,UniRep_F155,UniRep_F152,BiLSTM_F2987,BiLSTM_F2991

prediction/preprocessing/alphabets.py

+from __future__ import print_function, division
+
+import numpy as np
+
+class Alphabet:
+    def __init__(self, chars, encoding=None, mask=False, missing=255):
+        self.chars = np.frombuffer(chars, dtype=np.uint8)
+        self.encoding = np.zeros(256, dtype=np.uint8) + missing
+        if encoding is None:
+            self.encoding[self.chars] = np.arange(len(self.chars))
+            self.size = len(self.chars)
+        else:
+            self.encoding[self.chars] = encoding
+            self.size = encoding.max() + 1
+        self.mask = mask
+        if mask:
+            self.size -= 1
+
+    def __len__(self):
+        return self.size
+
+    def __getitem__(self, i):
+        return chr(self.chars[i])
+
+    def encode(self, x): 
+        """ encode a byte string into alphabet indices """
+        x = np.frombuffer(x, dtype=np.uint8)
+        return self.encoding[x]
+
+    def decode(self, x):
+        """ decode index array, x, to byte string of this alphabet """
+        string = self.chars[x]
+        return string.tobytes()
+
+    def unpack(self, h, k):
+        """ unpack integer h into array of this alphabet with length k """
+        n = self.size
+        kmer = np.zeros(k, dtype=np.uint8)
+        for i in reversed(range(k)):
+            c = h % n
+            kmer[i] = c
+            h = h // n
+        return kmer
+
+    def get_kmer(self, h, k):
+        """ retrieve byte string of length k decoded from integer h """
+        kmer = self.unpack(h, k)
+        return self.decode(kmer)
+        
+DNA = Alphabet(b'ACGT')
+        
+class Uniprot21(Alphabet):
+    def __init__(self, mask=False):
+        chars = alphabet = b'ARNDCQEGHILKMFPSTWYVXOUBZ'
+        encoding = np.arange(len(chars))
+        encoding[21:] = [11,4,20,20] # encode 'OUBZ' as synonyms
+        super(Uniprot21, self).__init__(chars, encoding=encoding, mask=mask, missing=20)
+
+class SDM12(Alphabet):
+    """
+    A D KER N TSQ YF LIVM C W H G P
+
+    See https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2732308/#B33
+    "Reduced amino acid alphabets exhibit an improved sensitivity and selectivity in fold assignment"
+    Peterson et al. 2009. Bioinformatics.
+    """
+    def __init__(self, mask=False):
+        chars = alphabet = b'ADKNTYLCWHGPXERSQFIVMOUBZ'
+        groups = [b'A',b'D',b'KERO',b'N',b'TSQ',b'YF',b'LIVM',b'CU',b'W',b'H',b'G',b'P',b'XBZ']
+        groups = {c:i for i in range(len(groups)) for c in groups[i]}
+        encoding = np.array([groups[c] for c in chars])
+        super(SDM12, self).__init__(chars, encoding=encoding, mask=mask)
+
+SecStr8 = Alphabet(b'HBEGITS ')
+
+
+

prediction/src/.idea/.gitignore

+# Default ignored files
+/shelf/
+/workspace.xml

prediction/src/.idea/inspectionProfiles/profiles_settings.xml

+<component name="InspectionProjectProfileManager">
+  <settings>
+    <option name="USE_PROJECT_PROFILE" value="false" />
+    <version value="1.0" />
+  </settings>
+</component>
\ No newline at end of file

prediction/src/.idea/misc.xml

+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectRootManager" version="2" project-jdk-name="Python 3.8" project-jdk-type="Python SDK" />
+</project>
\ No newline at end of file

prediction/src/.idea/modules.xml

+<?xml version="1.0" encoding="UTF-8"?>
+<project version="4">
+  <component name="ProjectModuleManager">
+    <modules>
+      <module fileurl="file://$PROJECT_DIR$/.idea/src.iml" filepath="$PROJECT_DIR$/.idea/src.iml" />
+    </modules>
+  </component>
+</project>
\ No newline at end of file

prediction/src/.idea/src.iml

+<?xml version="1.0" encoding="UTF-8"?>
+<module type="PYTHON_MODULE" version="4">
+  <component name="NewModuleRootManager">
+    <content url="file://$MODULE_DIR$" />
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>
\ No newline at end of file

prediction/src/SSA_embedding.py

+from __future__ import print_function,division
+
+import sys
+import os
+#sys.path.append('./embedding_model/')
+sys.path.append(os.path.join(os.pardir, os.pardir))
+
+import pandas as pd
+import torch
+import torch.nn as nn
+import time
+import logging
+
+
+DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+logging.info(DEVICE)
+
+from preprocessing.alphabets import Uniprot21
+ 
+import warnings
+ 
+warnings.filterwarnings('ignore')
+
+def unstack_lstm(lstm):
+    device = next(iter(lstm.parameters())).device
+
+    in_size = lstm.input_size
+    hidden_dim = lstm.hidden_size
+    layers = []
+    for i in range(lstm.num_layers):
+        layer = nn.LSTM(in_size, hidden_dim, batch_first=True, bidirectional=True)
+        layer.to(device)
+
+        attributes = ['weight_ih_l', 'weight_hh_l', 'bias_ih_l', 'bias_hh_l']
+        for attr in attributes:
+            dest = attr + '0'
+            src = attr + str(i)
+            getattr(layer, dest).data[:] = getattr(lstm, src)
+             
+            dest = attr + '0_reverse'
+            src = attr + str(i) + '_reverse'
+            getattr(layer, dest).data[:] = getattr(lstm, src)
+             
+        layer.flatten_parameters()
+        layers.append(layer)
+        in_size = 2*hidden_dim
+    return layers
+def embed_stack(x, lm_embed, lstm_stack, proj, include_lm=True, final_only=False):
+    zs = []
+    
+    x_onehot = x.new(x.size(0),x.size(1), 21).float().zero_()
+    x_onehot.scatter_(2,x.unsqueeze(2),1)
+    zs.append(x_onehot)
+    
+    h = lm_embed(x)
+    if include_lm and not final_only:
+        zs.append(h)
+
+    if lstm_stack is not None:
+        for lstm in lstm_stack:
+            h,_ = lstm(h)
+            if not final_only:
+                zs.append(h)
+    if proj is not None:
+        h = proj(h.squeeze(0)).unsqueeze(0)
+        zs.append(h)
+
+    z = torch.cat(zs, 2)
+    return z
+def embed_sequence(x, lm_embed, lstm_stack, proj, include_lm=True, final_only=False
+                  ,  pool='none', use_cuda=False):
+
+    if len(x) == 0:
+        return None
+
+    alphabet = Uniprot21()
+    x = x.upper()
+    x = alphabet.encode(x)
+    x = torch.from_numpy(x)
+    if use_cuda:
+        x = x.to(DEVICE)
+    with torch.no_grad():
+        x = x.long().unsqueeze(0)
+        z = embed_stack(x, lm_embed, lstm_stack, proj
+                       , include_lm=include_lm, final_only=final_only)
+        z = z.squeeze(0)
+        if pool == 'sum':
+            z = z.sum(0)
+        elif pool == 'max':
+            z,_ = z.max(0)
+        elif pool == 'avg':
+            z = z.mean(0)
+        z = z.cpu().numpy()
+
+    return z
+def load_model(path, use_cuda=False):
+     
+    encoder = torch.load(path)
+    encoder.eval()
+    if use_cuda:
+        encoder=encoder.to(DEVICE)
+    encoder = encoder.embedding
+
+    lm_embed = encoder.embed
+    lstm_stack = unstack_lstm(encoder.rnn)
+    proj = encoder.proj
+
+    return lm_embed, lstm_stack, proj
+
+
+def SSA_Embed(input_seq):
+    T0 = time.time()
+    SSAEMB_ = []
+    PID_ =[]
+    # inData = fasta.fasta2csv(fastaFile)
+    # SEQ_ = inData["Seq"]
+    # PID_ = inData["PID"]
+    # CLASS_ = inData["Class"]
+
+    logging.info("SSA Embedding...")
+    print("Loading SSA Model...", file=sys.stderr, end='\r')
+    lm_embed, lstm_stack, proj = load_model("./prediction/embbed_models/SSA_embed.model", use_cuda=True)
+
+    include_lm = True
+    final_only = True
+    for key, value in input_seq.items():
+        PID_.append(key)
+        sequence = str(value).encode("utf-8")
+        z = embed_sequence(sequence, lm_embed, lstm_stack, proj
+                           , include_lm=include_lm, final_only=final_only
+                           , pool='avg', use_cuda=True)
+
+        SSAEMB_.append(z)
+    # count += 1
+    # print('#{}$'.format(count), file=sys.stderr, end='\r')
+
+    ssa_feature = pd.DataFrame(SSAEMB_)
+    col = ["SSA_F" + str(i + 1) for i in range(0, 121)]
+    ssa_feature.columns = col
+    ssa_feature = pd.concat([ssa_feature], axis=1)
+    ssa_feature.index = PID_
+    logging.info(ssa_feature.shape)
+    # ssa_feature.to_csv("./feature_vectors/SSA_feature.csv")
+    logging.info("SSA embedding finished@￥￥￥￥￥")
+    logging.info("it took %0.3f mins.\n" % ((time.time() - T0) / 60))
+
+    return ssa_feature
+
+def BiLSTM_Embed(input_seq):
+    T0=time.time()
+    BiLSTMEMB_=[]
+    PID=[]
+    logging.info("\nBiLSTM Embedding...")
+    print("Loading BiLSTM Model...", file=sys.stderr, end='\r')
+    lm_embed, lstm_stack, proj = load_model("./prediction/embbed_models/SSA_embed.model", use_cuda=True)
+    proj = None
+
+    for key,value in input_seq.items():
+        PID.append(key)
+        sequence = str(value).encode("utf-8")
+        z = embed_sequence(sequence, lm_embed, lstm_stack, proj
+                                  ,  final_only=False,include_lm = True
+                                  , pool='avg', use_cuda=True)
+        BiLSTMEMB_.append(z)
+    bilstm_feature=pd.DataFrame(BiLSTMEMB_)
+    col=["BiLSTM_F"+str(i+1) for i in range(0,3605)]
+    bilstm_feature.columns=col
+    bilstm_feature=pd.concat([bilstm_feature],axis=1)
+    bilstm_feature.index=PID
+    # bilstm_feature.to_csv("./feature_vectors/bilstm_feature.csv")
+    logging.info("BiLSTM embedding finished@@￥￥￥￥￥")
+    logging.info("it took %0.3f mins.\n"%((time.time()-T0)/60))
+
+    return bilstm_feature
+
+    
+    
+     
+    
+
+
+
+

prediction/src/UniRep_emb.py

+from __future__ import print_function,division
+
+import sys
+import os
+sys.path.append(os.pardir)
+sys.path.append(os.path.join(os.pardir, os.pardir))
+ 
+import time
+import pandas as pd
+import torch
+import warnings
+ 
+warnings.filterwarnings('ignore')
+DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+from tape import UniRepModel,TAPETokenizer
+
+def UniRep_Embed(input_seq):
+    T0=time.time()
+    UNIREPEB_=[]
+    PID = []
+    print("UniRep Embedding...")
+
+    model = UniRepModel.from_pretrained('babbler-1900')
+    model=model.to(DEVICE)
+    tokenizer = TAPETokenizer(vocab='unirep')
+
+    for key,value in input_seq.items():
+        PID.append(key)
+        sequence = value
+        if len(sequence) == 0:
+            print('# WARNING: sequence', PID, 'has length=0. Skipping.', file=sys.stderr)
+            continue
+        with torch.no_grad():
+            token_ids = torch.tensor([tokenizer.encode(sequence)])
+            token_ids = token_ids.to(DEVICE)
+            output = model(token_ids)
+            unirep_output = output[0]
+            unirep_output=torch.squeeze(unirep_output)
+            unirep_output= unirep_output.mean(0)
+            unirep_output = unirep_output.cpu().numpy()
+            UNIREPEB_.append(unirep_output.tolist())
+    unirep_feature=pd.DataFrame(UNIREPEB_)
+    col=["UniRep_F"+str(i+1) for i in range(0,1900)]
+    unirep_feature.columns=col
+    unirep_feature=pd.concat([unirep_feature],axis=1)
+    unirep_feature.index=PID
+    # print(unirep_feature.shape)
+    unirep_feature.to_csv("./dataset/unirep_feature.csv")
+    print("Getting Deep Representation Learning Features with UniRep is done.")
+    print("it took %0.3f mins.\n"%((time.time()-T0)/60))
+
+    return unirep_feature
+    
\ No newline at end of file

prediction/src/models/comparison.py

+from __future__ import print_function,division
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+
+class L1(nn.Module):
+    def forward(self, x, y):
+        return -torch.sum(torch.abs(x.unsqueeze(1)-y), -1)
+
+
+class L2(nn.Module):
+    def forward(self, x, y):
+        return -torch.sum((x.unsqueeze(1)-y)**2, -1)
+
+
+class DotProduct(nn.Module):
+    def forward(self, x, y):
+        return torch.mm(x, y.t())
+
+
+def pad_gap_scores(s, gap):
+    col = gap.expand(s.size(0), 1)
+    s = torch.cat([s, col], 1)
+    row = gap.expand(1, s.size(1))
+    s = torch.cat([s, row], 0)
+    return s
+
+
+class OrdinalRegression(nn.Module):
+    def __init__(self, embedding, n_classes, compare=L1()
+                , align_method='ssa', beta_init=10
+                , allow_insertions=False, gap_init=-10
+                ):
+        super(OrdinalRegression, self).__init__()
+        
+        self.embedding = embedding
+        self.n_out = n_classes
+
+        self.compare = compare
+        self.align_method = align_method
+        self.allow_insertions = allow_insertions
+        self.gap = nn.Parameter(torch.FloatTensor([gap_init]))
+
+        self.theta = nn.Parameter(torch.ones(1,n_classes-1))
+        self.beta = nn.Parameter(torch.zeros(n_classes-1)+beta_init)
+        self.clip()
+
+    def forward(self, x):
+        return self.embedding(x)
+
+    def clip(self):
+        # clip the weights of ordinal regression to be non-negative
+        self.theta.data.clamp_(min=0)
+
+    def score(self, z_x, z_y):
+
+        if self.align_method == 'ssa':
+            s = self.compare(z_x, z_y)
+            if self.allow_insertions:
+                s = pad_gap_scores(s, self.gap)
+
+            a = F.softmax(s, 1)
+            b = F.softmax(s, 0)
+
+            if self.allow_insertions:
+                index = s.size(0)-1
+                index = s.data.new(1).long().fill_(index)
+                a = a.index_fill(0, index, 0)
+
+                index = s.size(1)-1
+                index = s.data.new(1).long().fill_(index)
+                b = b.index_fill(1, index, 0)
+
+            a = a + b - a*b
+            c = torch.sum(a*s)/torch.sum(a)
+
+        elif self.align_method == 'ua':
+            s = self.compare(z_x, z_y)
+            c = torch.mean(s)
+
+        elif self.align_method == 'me':
+            z_x = z_x.mean(0)
+            z_y = z_y.mean(0)
+            c = self.compare(z_x.unsqueeze(0), z_y.unsqueeze(0)).squeeze(0)
+
+        else:
+            raise Exception('Unknown alignment method: ' + self.align_method)
+
+        logits = c*self.theta + self.beta
+        return logits.view(-1)
+
+
+

prediction/src/models/embedding.py

+from __future__ import print_function,division
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from torch.nn.utils.rnn import PackedSequence
+
+
+class LMEmbed(nn.Module):
+    def __init__(self, nin, nout, lm, padding_idx=-1, transform=nn.ReLU()
+                , sparse=False):
+        super(LMEmbed, self).__init__()
+
+        if padding_idx == -1:
+            padding_idx = nin-1
+
+        self.lm = lm
+        self.embed = nn.Embedding(nin, nout, padding_idx=padding_idx, sparse=sparse)
+        self.proj = nn.Linear(lm.hidden_size(), nout)
+        self.transform = transform
+        self.nout = nout
+
+    def forward(self, x):
+        packed = type(x) is PackedSequence
+        h_lm = self.lm.encode(x)
+
+        # embed and unpack if packed
+        if packed:
+            h = self.embed(x.data)
+            h_lm = h_lm.data
+        else:
+            h = self.embed(x)
+
+        # project
+        h_lm = self.proj(h_lm)
+        h = self.transform(h + h_lm)
+
+        # repack if needed
+        if packed:
+            h = PackedSequence(h, x.batch_sizes)
+
+        return h
+
+
+class Linear(nn.Module):
+    def __init__(self, nin, nhidden, nout, padding_idx=-1,
+                 sparse=False, lm=None):
+        super(Linear, self).__init__()
+        
+        if padding_idx == -1:
+            padding_idx = nin-1
+        
+        if lm is not None:
+            self.embed = LMEmbed(nin, nhidden, lm, padding_idx=padding_idx, sparse=sparse)
+            self.proj = nn.Linear(self.embed.nout, nout)
+            self.lm = True
+        else:
+            self.proj = nn.Embedding(nin, nout, padding_idx=padding_idx, sparse=sparse)
+            self.lm = False
+
+        self.nout = nout
+        
+        
+    def forward(self, x):
+        
+        if self.lm:
+            h = self.embed(x)
+            if type(h) is PackedSequence:
+                h = h.data
+                z = self.proj(h)
+                z = PackedSequence(z, x.batch_sizes)
+            else:
+                h = h.view(-1, h.size(2))
+                z = self.proj(h)
+                z = z.view(x.size(0), x.size(1), -1)
+        else:
+            if type(x) is PackedSequence:
+                z = self.embed(x.data)
+                z = PackedSequence(z, x.batch_sizes)
+            else:
+                z = self.embed(x)
+    
+        return z
+
+
+class StackedRNN(nn.Module):
+    def __init__(self, nin, nembed, nunits, nout, nlayers=2, padding_idx=-1, dropout=0,
+                 rnn_type='lstm', sparse=False, lm=None):
+        super(StackedRNN, self).__init__()
+        
+        if padding_idx == -1:
+            padding_idx = nin-1
+        
+        if lm is not None:
+            self.embed = LMEmbed(nin, nembed, lm, padding_idx=padding_idx, sparse=sparse)
+            nembed = self.embed.nout
+            self.lm = True
+        else:
+            self.embed = nn.Embedding(nin, nembed, padding_idx=padding_idx, sparse=sparse)
+            self.lm = False
+
+        if rnn_type == 'lstm':
+            RNN = nn.LSTM
+        elif rnn_type == 'gru':
+            RNN = nn.GRU
+
+        self.dropout = nn.Dropout(p=dropout)
+        if nlayers == 1:
+            dropout = 0
+
+        self.rnn = RNN(nembed, nunits, nlayers, batch_first=True
+                      , bidirectional=True, dropout=dropout)
+        self.proj = nn.Linear(2*nunits, nout)
+        self.nout = nout
+
+        
+        
+    def forward(self, x):
+        
+        if self.lm:
+            h = self.embed(x)
+        else:
+            if type(x) is PackedSequence:
+                h = self.embed(x.data)
+                h = PackedSequence(h, x.batch_sizes)
+            else:
+                h = self.embed(x)
+            
+        h,_ = self.rnn(h)
+        
+        if type(h) is PackedSequence:
+            h = h.data
+            h = self.dropout(h)
+            z = self.proj(h)
+            z = PackedSequence(z, x.batch_sizes)
+        else:
+            h = h.view(-1, h.size(2))
+            h = self.dropout(h)
+            z = self.proj(h)
+            z = z.view(x.size(0), x.size(1), -1)
+    
+        return z
+
+

prediction/src/models/multitask.py

+from __future__ import print_function,division
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+from .comparison import L1, pad_gap_scores
+
+
+class SCOPCM(nn.Module):
+    def __init__(self, embedding, similarity_kwargs={},
+                 cmap_kwargs={}):
+        super(SCOPCM, self).__init__()
+
+        self.embedding = embedding
+        embed_dim = embedding.nout
+
+        self.scop_predict = OrdinalRegression(5, **similarity_kwargs)
+        self.cmap_predict = ConvContactMap(embed_dim, **cmap_kwargs)
+
+    def clip(self):
+        self.scop_predict.clip()
+        self.cmap_predict.clip()
+
+    def forward(self, x):
+        return self.embedding(x)
+
+    def score(self, z_x, z_y):
+        return self.scop_predict(z_x, z_y)
+
+    def predict(self, z):
+        return self.cmap_predict(z)
+
+
+class ConvContactMap(nn.Module):
+    def __init__(self, embed_dim, hidden_dim=50, width=7, act=nn.ReLU()):
+        super(ConvContactMap, self).__init__()
+        self.hidden = nn.Conv2d(2*embed_dim, hidden_dim, 1)
+        self.act = act
+        self.conv = nn.Conv2d(hidden_dim, 1, width, padding=width//2)
+        self.clip()
+
+    def clip(self):
+        # force the conv layer to be transpose invariant
+        w = self.conv.weight
+        self.conv.weight.data[:] = 0.5*(w + w.transpose(2,3))
+
+    def forward(self, z):
+        return self.predict(z)
+
+    def predict(self, z):
+        # z is (b,L,d)
+        z = z.transpose(1, 2) # (b,d,L)
+        z_dif = torch.abs(z.unsqueeze(2) - z.unsqueeze(3))
+        z_mul = z.unsqueeze(2)*z.unsqueeze(3)
+        z = torch.cat([z_dif, z_mul], 1)
+        # (b,2d,L,L)
+        h = self.act(self.hidden(z))
+        logits = self.conv(h).squeeze(1)
+        return logits
+    
+
+class OrdinalRegression(nn.Module):
+    def __init__(self, n_classes, compare=L1()
+                , align_method='ssa', beta_init=10
+                , allow_insertions=False, gap_init=-10
+                ):
+        super(OrdinalRegression, self).__init__()
+        
+        self.n_out = n_classes
+
+        self.compare = compare
+        self.align_method = align_method
+        self.allow_insertions = allow_insertions
+        self.gap = nn.Parameter(torch.FloatTensor([gap_init]))
+
+        self.theta = nn.Parameter(torch.ones(1,n_classes-1))
+        self.beta = nn.Parameter(torch.zeros(n_classes-1)+beta_init)
+        self.clip()
+
+    def clip(self):
+        # clip the weights of ordinal regression to be non-negative
+        self.theta.data.clamp_(min=0)
+
+    def forward(self, z_x, z_y):
+
+        if self.align_method == 'ssa':
+            s = self.compare(z_x, z_y)
+            if self.allow_insertions:
+                s = pad_gap_scores(s, self.gap)
+
+            a = F.softmax(s, 1)
+            b = F.softmax(s, 0)
+
+            if self.allow_insertions:
+                index = s.size(0)-1
+                index = s.data.new(1).long().fill_(index)
+                a = a.index_fill(0, index, 0)
+
+                index = s.size(1)-1
+                index = s.data.new(1).long().fill_(index)
+                b = b.index_fill(1, index, 0)
+
+            a = a + b - a*b
+            c = torch.sum(a*s)/torch.sum(a)
+
+        elif self.align_method == 'ua':
+            s = self.compare(z_x, z_y)
+            c = torch.mean(s)
+
+        elif self.align_method == 'me':
+            z_x = z_x.mean(0)
+            z_y = z_y.mean(0)
+            c = self.compare(z_x.unsqueeze(0), z_y.unsqueeze(0)).squeeze(0)
+
+        else:
+            raise Exception('Unknown alignment method: ' + self.align_method)
+
+        logits = c*self.theta + self.beta
+        return logits.view(-1)
+
+
+
+

prediction/src/models/sequence.py

+from __future__ import print_function,division
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from torch.nn.utils.rnn import PackedSequence, pack_padded_sequence, pad_packed_sequence
+
+class BiLM(nn.Module):
+    def __init__(self, nin, nout, embedding_dim, hidden_dim, num_layers
+                , tied=True, mask_idx=None, dropout=0):
+        super(BiLM, self).__init__()
+        
+        if mask_idx is None:
+            mask_idx = nin-1
+        self.mask_idx = mask_idx
+        self.embed = nn.Embedding(nin, embedding_dim, padding_idx=mask_idx)
+        self.dropout = nn.Dropout(p=dropout)
+
+        self.tied = tied
+        if tied:
+            layers = []
+            nin = embedding_dim
+            for _ in range(num_layers):
+                layers.append(nn.LSTM(nin, hidden_dim, 1, batch_first=True))
+                nin = hidden_dim
+            self.rnn = nn.ModuleList(layers)
+        else:
+            layers = []
+            nin = embedding_dim
+            for _ in range(num_layers):
+                layers.append(nn.LSTM(nin, hidden_dim, 1, batch_first=True))
+                nin = hidden_dim
+            self.lrnn = nn.ModuleList(layers)
+
+            layers = []
+            nin = embedding_dim
+            for _ in range(num_layers):
+                layers.append(nn.LSTM(nin, hidden_dim, 1, batch_first=True))
+                nin = hidden_dim
+            self.rrnn = nn.ModuleList(layers)
+
+        self.linear = nn.Linear(hidden_dim, nout)
+
+    def hidden_size(self):
+        h = 0
+        if self.tied:
+            for layer in self.rnn:
+                h += 2*layer.hidden_size
+        else:
+            for layer in self.lrnn:
+                h += layer.hidden_size
+            for layer in self.rrnn:
+                h += layer.hidden_size
+        return h
+
+    
+    def reverse(self, h):
+        packed = type(h) is PackedSequence
+        if packed:
+            h,batch_sizes = pad_packed_sequence(h, batch_first=True)
+            h_rvs = h.clone().zero_()
+            for i in range(h.size(0)):
+                n = batch_sizes[i]
+                idx = [j for j in range(n-1, -1, -1)]
+                idx = torch.LongTensor(idx).to(h.device)
+                h_rvs[i,:n] = h[i].index_select(0, idx)
+            # repack h_rvs
+            h_rvs = pack_padded_sequence(h_rvs, batch_sizes, batch_first=True)
+        else:
+            idx = [i for i in range(h.size(1)-1, -1, -1)]
+            idx = torch.LongTensor(idx).to(h.device)
+            h_rvs = h.index_select(1, idx)
+        return h_rvs
+
+
+    def transform(self, z_fwd, z_rvs, last_only=False):
+        # sequences are flanked by the start/stop token as:
+        # [stop, x, stop]
+
+        # z_fwd should be [stop,x]
+        # z_rvs should be [x,stop] reversed
+
+        # first, do the forward direction
+        if self.tied:
+            layers = self.rnn
+        else:
+            layers = self.lrnn
+
+        h_fwd = []
+        h = z_fwd
+        for rnn in layers:
+            h,_ = rnn(h)
+            if type(h) is PackedSequence:
+                h = PackedSequence(self.dropout(h.data), h.batch_sizes)
+            else:
+                h = self.dropout(h)
+            if not last_only:
+                h_fwd.append(h)
+        if last_only:
+            h_fwd = h
+
+        # now, do the reverse direction
+        if self.tied:
+            layers = self.rnn
+        else:
+            layers = self.rrnn
+
+        # we'll need to reverse the direction of these
+        # hidden states back to match forward direction
+
+        h_rvs = []
+        h = z_rvs
+        for rnn in layers:
+            h,_ = rnn(h)
+            if type(h) is PackedSequence:
+                h = PackedSequence(self.dropout(h.data), h.batch_sizes)
+            else:
+                h = self.dropout(h)
+            if not last_only:
+                h_rvs.append(self.reverse(h))
+        if last_only:
+            h_rvs = self.reverse(h)
+
+        return h_fwd,h_rvs
+
+
+    def embed_and_split(self, x, pad=False):
+
+        packed = type(x) is PackedSequence
+        if packed:
+            x,batch_sizes = pad_packed_sequence(x, batch_first=True)
+
+        if pad:
+            # pad x with the start/stop token
+            x = x + 1
+            ## append start/stop tokens to x
+            x_ = x.data.new(x.size(0), x.size(1)+2).zero_()
+            if packed:
+                for i in range(len(batch_sizes)):
+                    n = batch_sizes[i]
+                    x_[i,1:n+1] = x[i,:n]
+                batch_sizes = [s+2 for s in batch_sizes]
+            else:
+                x_[:,1:-1] = x
+            x = x_
+
+        # sequences x are flanked by the start/stop token as:
+        # [stop, x, stop]
+
+        # now, encode x as distributed vectors
+        z = self.embed(x)
+
+        # to pass to transform, we discard the last element for z_fwd and the first element for z_rvs
+        z_fwd = z[:,:-1]
+        z_rvs = z[:,1:]
+        if packed:
+            lengths = [s-1 for s in batch_sizes]
+            z_fwd = pack_padded_sequence(z_fwd, lengths, batch_first=True)
+            z_rvs = pack_padded_sequence(z_rvs, lengths, batch_first=True)
+        # reverse z_rvs
+        z_rvs = self.reverse(z_rvs)
+
+        return z_fwd, z_rvs
+
+
+    def encode(self, x):
+        z_fwd,z_rvs = self.embed_and_split(x, pad=True)
+        h_fwd_layers,h_rvs_layers = self.transform(z_fwd, z_rvs)
+
+        # concatenate hidden layers together
+        packed = type(z_fwd) is PackedSequence
+        concat = []
+        for h_fwd,h_rvs in zip(h_fwd_layers,h_rvs_layers):
+            if packed:
+                h_fwd,batch_sizes = pad_packed_sequence(h_fwd, batch_first=True)
+                h_rvs,batch_sizes = pad_packed_sequence(h_rvs, batch_first=True)
+            # discard last element of h_fwd and first element of h_rvs
+            h_fwd = h_fwd[:,:-1]
+            h_rvs = h_rvs[:,1:]
+            
+            # accumulate for concatenation
+            concat.append(h_fwd)
+            concat.append(h_rvs)
+        
+        h = torch.cat(concat, 2)
+        if packed:
+            batch_sizes = [s-1 for s in batch_sizes]
+            h = pack_padded_sequence(h, batch_sizes, batch_first=True)
+
+        return h 
+
+
+    def forward(self, x):
+        # x's are already flanked by the star/stop token as:
+        # [stop, x, stop]
+        z_fwd,z_rvs = self.embed_and_split(x, pad=False)
+
+        h_fwd,h_rvs = self.transform(z_fwd, z_rvs, last_only=True)
+
+        packed = type(z_fwd) is PackedSequence
+        if packed:
+            h_flat = h_fwd.data
+            logp_fwd = self.linear(h_flat)
+            logp_fwd = PackedSequence(logp_fwd, h_fwd.batch_sizes)
+
+            h_flat = h_rvs.data
+            logp_rvs = self.linear(h_flat)
+            logp_rvs = PackedSequence(logp_rvs, h_rvs.batch_sizes)
+
+            logp_fwd,batch_sizes = pad_packed_sequence(logp_fwd, batch_first=True)
+            logp_rvs,batch_sizes = pad_packed_sequence(logp_rvs, batch_first=True)
+
+        else:
+            b = h_fwd.size(0)
+            n = h_fwd.size(1)
+            h_flat = h_fwd.contiguous().view(-1, h_fwd.size(2))
+            logp_fwd = self.linear(h_flat)
+            logp_fwd = logp_fwd.view(b, n, -1)
+
+            h_flat = h_rvs.contiguous().view(-1, h_rvs.size(2))
+            logp_rvs = self.linear(h_flat)
+            logp_rvs = logp_rvs.view(b, n, -1)
+
+        # prepend forward logp with zero
+        # postpend reverse logp with zero
+
+        b = h_fwd.size(0)
+        zero = h_fwd.data.new(b,1,logp_fwd.size(2)).zero_()
+        logp_fwd = torch.cat([zero, logp_fwd], 1)
+        logp_rvs = torch.cat([logp_rvs, zero], 1)
+
+        logp = F.log_softmax(logp_fwd + logp_rvs, dim=2)
+
+        if packed:
+            batch_sizes = [s+1 for s in batch_sizes]
+            logp = pack_padded_sequence(logp, batch_sizes, batch_first=True)
+
+        return logp 
+
+
+

prediction/train_label.csv

+PID,Category
+negative38|0,0
+negative135|0,0
+negative243|0,0
+negative264|0,0
+negative36|0,0
+negative162|0,0
+negative19|0,0
+negative221|0,0
+negative113|0,0
+negative61|0,0
+negative277|0,0
+negative66|0,0
+negative75|0,0
+negative184|0,0
+negative186|0,0
+negative115|0,0
+negative276|0,0
+negative6|0,0
+negative142|0,0
+negative12|0,0
+negative82|0,0
+negative171|0,0
+negative303|0,0
+negative181|0,0
+negative20|0,0
+negative241|0,0
+negative76|0,0
+negative260|0,0
+negative57|0,0
+negative9|0,0
+negative25|0,0
+negative50|0,0
+negative198|0,0
+negative267|0,0
+negative218|0,0
+negative33|0,0
+negative262|0,0
+negative169|0,0
+negative132|0,0
+negative252|0,0
+negative46|0,0
+negative284|0,0
+negative291|0,0
+negative253|0,0
+negative14|0,0
+negative84|0,0
+negative35|0,0
+negative62|0,0
+negative119|0,0
+negative232|0,0
+negative182|0,0
+negative295|0,0
+negative190|0,0
+negative164|0,0
+negative296|0,0
+negative167|0,0
+negative45|0,0
+negative211|0,0
+negative189|0,0
+negative143|0,0
+negative287|0,0
+negative161|0,0
+negative230|0,0
+negative178|0,0
+negative3|0,0
+negative271|0,0
+negative70|0,0
+negative107|0,0
+negative259|0,0
+negative204|0,0
+negative85|0,0
+negative172|0,0
+negative7|0,0
+negative157|0,0
+negative129|0,0
+negative31|0,0
+negative149|0,0
+negative128|0,0
+negative16|0,0
+negative175|0,0
+negative248|0,0
+negative265|0,0
+negative43|0,0
+negative109|0,0
+negative44|0,0
+negative205|0,0
+negative200|0,0
+negative56|0,0
+negative240|0,0
+negative140|0,0
+negative133|0,0
+negative275|0,0
+negative247|0,0
+negative242|0,0
+negative91|0,0
+negative92|0,0
+negative281|0,0
+negative166|0,0
+negative250|0,0
+negative199|0,0
+negative59|0,0
+negative18|0,0
+negative225|0,0
+negative69|0,0
+negative90|0,0
+negative64|0,0
+negative301|0,0
+negative206|0,0
+negative95|0,0
+negative147|0,0
+negative210|0,0
+negative2|0,0
+negative106|0,0
+negative183|0,0
+negative197|0,0
+negative236|0,0
+negative78|0,0
+negative269|0,0
+negative215|0,0
+negative134|0,0
+negative302|0,0
+negative188|0,0
+negative233|0,0
+negative153|0,0
+negative258|0,0
+negative48|0,0
+negative68|0,0
+negative27|0,0
+negative194|0,0
+negative224|0,0
+negative104|0,0
+negative65|0,0
+negative17|0,0
+negative136|0,0
+negative251|0,0
+negative63|0,0
+negative165|0,0
+negative229|0,0
+negative257|0,0
+negative280|0,0
+negative152|0,0
+negative77|0,0
+negative49|0,0
+negative168|0,0
+negative89|0,0
+negative117|0,0
+negative79|0,0
+negative130|0,0
+negative58|0,0
+negative288|0,0
+negative102|0,0
+negative144|0,0
+negative219|0,0
+negative41|0,0
+negative173|0,0
+negative237|0,0
+negative86|0,0
+negative244|0,0
+negative116|0,0
+negative254|0,0
+negative30|0,0
+negative15|0,0
+negative123|0,0
+negative101|0,0
+negative40|0,0
+negative97|0,0
+negative174|0,0
+negative238|0,0
+negative74|0,0
+negative293|0,0
+negative23|0,0
+negative81|0,0
+negative5|0,0
+negative120|0,0
+negative87|0,0
+negative151|0,0
+negative297|0,0
+negative227|0,0
+negative146|0,0
+negative47|0,0
+negative125|0,0
+negative114|0,0
+negative282|0,0
+negative179|0,0
+negative4|0,0
+negative246|0,0
+negative298|0,0
+negative239|0,0
+negative268|0,0
+negative100|0,0
+negative304|0,0
+negative155|0,0
+negative28|0,0
+negative131|0,0
+negative94|0,0
+negative283|0,0
+negative158|0,0
+negative8|0,0
+negative261|0,0
+negative176|0,0
+negative208|0,0
+negative83|0,0
+negative124|0,0
+negative279|0,0
+negative170|0,0
+negative273|0,0
+negative93|0,0
+negative255|0,0
+negative207|0,0
+negative223|0,0
+negative24|0,0
+negative145|0,0
+negative122|0,0
+negative99|0,0
+negative220|0,0
+negative196|0,0
+negative154|0,0
+negative29|0,0
+negative228|0,0
+negative292|0,0
+negative71|0,0
+negative105|0,0
+negative278|0,0
+negative202|0,0
+negative187|0,0
+negative60|0,0
+negative103|0,0
+negative26|0,0
+negative177|0,0
+negative201|0,0
+negative159|0,0
+negative192|0,0
+negative156|0,0
+negative266|0,0
+negative13|0,0
+negative98|0,0
+negative39|0,0
+negative22|0,0
+negative234|0,0
+negative226|0,0
+negative185|0,0
+negative121|0,0
+negative212|0,0
+positive29|1,1
+positive69|1,1
+positive4|1,1
+positive139|1,1
+positive123|1,1
+positive21|1,1
+positive117|1,1
+positive86|1,1
+positive109|1,1
+positive24|1,1
+positive76|1,1
+positive30|1,1
+positive72|1,1
+positive77|1,1
+positive7|1,1
+positive133|1,1
+positive145|1,1
+positive147|1,1
+positive34|1,1
+positive148|1,1
+positive82|1,1
+positive115|1,1
+positive136|1,1
+positive33|1,1
+positive132|1,1
+positive78|1,1
+positive46|1,1
+positive121|1,1
+positive8|1,1
+positive108|1,1
+positive90|1,1
+positive114|1,1
+positive146|1,1
+positive57|1,1
+positive95|1,1
+positive106|1,1
+positive141|1,1
+positive41|1,1
+positive50|1,1
+positive14|1,1
+positive74|1,1
+positive100|1,1
+positive10|1,1
+positive66|1,1
+positive94|1,1
+positive45|1,1
+positive137|1,1
+positive36|1,1
+positive35|1,1
+positive64|1,1
+positive87|1,1
+positive142|1,1
+positive84|1,1
+positive5|1,1
+positive28|1,1
+positive128|1,1
+positive67|1,1
+positive62|1,1
+positive44|1,1
+positive138|1,1
+positive22|1,1
+positive103|1,1
+positive96|1,1
+positive104|1,1
+positive15|1,1
+positive99|1,1
+positive80|1,1
+positive17|1,1
+positive119|1,1
+positive43|1,1
+positive65|1,1
+positive113|1,1
+positive151|1,1
+positive49|1,1
+positive144|1,1
+positive11|1,1
+positive12|1,1
+positive122|1,1
+positive73|1,1
+positive54|1,1
+positive23|1,1
+positive13|1,1
+positive126|1,1
+positive97|1,1
+positive39|1,1
+positive25|1,1
+positive79|1,1
+positive83|1,1
+positive85|1,1
+positive26|1,1
+positive32|1,1
+positive92|1,1
+positive9|1,1
+positive3|1,1
+positive20|1,1
+positive134|1,1
+positive110|1,1
+positive102|1,1
+positive42|1,1
+positive60|1,1
+positive16|1,1
+positive58|1,1
+positive53|1,1
+positive31|1,1
+positive130|1,1
+positive93|1,1
+positive124|1,1
+positive71|1,1
+positive89|1,1
+positive51|1,1
+positive107|1,1
+positive37|1,1
+positive59|1,1
+positive150|1,1
+positive61|1,1
+positive27|1,1
+positive143|1,1
+positive19|1,1
+positive6|1,1
+positive105|1,1
+positive63|1,1

requirements.txt

+joblib==1.0.1
+tape_proteins==0.4
+torch==1.2.0+cpu
+numpy==1.19.2
+pandas==1.2.0
+Bio==0.4.1
+tape==1.0