Mercurial > repos > jay > pdaug_peptide_ngrams

comparison PDAUG_Sequence_Property_Based_Descriptors/PDAUG_Sequence_Property_Based_Descriptors.py @ 5:d7e684975db3 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/jaidevjoshi83/pdaug commit ac4353ca5c0ac9ce60df9f4bf160ed08b99fbee3"
author jay
date Thu, 28 Jan 2021 04:05:16 +0000
parents 7557b48b2872
children
comparison
equal deleted inserted replaced
4:cb2b359a6600 5:d7e684975db3
41 41
42 df = pd.DataFrame(lines) 42 df = pd.DataFrame(lines)
43 43
44 return df 44 return df
45 45
46 def Decriptor_generator(InFile, Lamda, Weight, DesType, Out_file): 46 def Decriptor_generator(infile, lamda, weight, maxlag, destype, out_file):
47 47
48 list_pep_name = [] 48 list_pep_name = []
49 f = open(InFile) 49 f = open(infile)
50 lines = f.readlines() 50 lines = f.readlines()
51 51
52 for line in lines: 52 for line in lines:
53 if ">" in line: 53 if ">" in line:
54 pass 54 pass
60 for seq in list_pep_name: 60 for seq in list_pep_name:
61 61
62 protein = PyPro() 62 protein = PyPro()
63 protein.ReadProteinSequence(seq) 63 protein.ReadProteinSequence(seq)
64 64
65 if DesType == 'PAAC': 65
66 DS = protein.GetPAAC(lamda=int(Lamda), weight=float(Weight)) 66 if destype == "GetAAComp":
67 DS = protein.GetAAComp()
67 df = pd.DataFrame(DS, index=[0]) 68 df = pd.DataFrame(DS, index=[0])
68 elif DesType == 'APAAC': 69 elif destype == "GetDPComp":
69 DS = protein.GetAPAAC(lamda=int(Lamda), weight=float(Weight)) 70 DS = protein.GetDPComp()
70 df = pd.DataFrame(DS, index=[0]) 71 df = pd.DataFrame(DS, index=[0])
71 elif DesType == 'CTD': 72 elif destype == "GetTPComp":
73 DS = protein.GetTPComp()
74 df = pd.DataFrame(DS, index=[0])
75 elif destype == "GetMoreauBrotoAuto":
76 DS = protein.GetMoreauBrotoAuto()
77 df = pd.DataFrame(DS, index=[0])
78 elif destype =="GetMoranAuto":
79 DS = protein.GetMoranAuto()
80 df = pd.DataFrame(DS, index=[0])
81 elif destype =="GetGearyAuto":
82 DS = protein.GetGearyAuto()
83 df = pd.DataFrame(DS, index=[0])
84 elif destype == "GetCTD":
72 DS = protein.GetCTD() 85 DS = protein.GetCTD()
73 df = pd.DataFrame(DS, index=[0]) 86 df = pd.DataFrame(DS, index=[0])
74 elif DesType == 'DPComp': 87 elif destype == "GetPAAC":
75 DS = protein.GetDPComp() 88 DS = protein.GetPAAC(lamda=int(lamda), weight=float(weight))
76 df = pd.DataFrame(DS, index=[0]) 89 df = pd.DataFrame(DS, index=[0])
77 elif DesType == 'GearyAuto': 90 elif destype == "GetAPAAC":
78 DS = protein.GetGearyAuto() 91 DS = protein.GetAPAAC(lamda=int(lamda), weight=float(weight))
79 df = pd.DataFrame(DS, index=[0]) 92 df = pd.DataFrame(DS, index=[0])
80 elif DesType == 'MoranAuto': 93 elif destype =="GetSOCN":
81 DS = protein.GetMoranAuto() 94 DS = protein.GetSOCN(maxlag=int(maxlag))
82 df = pd.DataFrame(DS, index=[0]) 95 df = pd.DataFrame(DS, index=[0])
83 elif DesType == 'MoreauBrotoAuto': 96 elif destype =="GetQSO":
84 DS = protein.GetMoreauBrotoAuto() 97 DS = protein.GetQSO(maxlag=int(maxlag), weight=float(weight))
85 df = pd.DataFrame(DS, index=[0]) 98 df = pd.DataFrame(DS, index=[0])
86 elif DesType == 'QSO': 99 elif destype == "GetTriad":
87 DS = protein.GetQSO() 100 DS = protein.GetTriad()
88 df = pd.DataFrame(DS, index=[0]) 101 df = pd.DataFrame(DS, index=[0])
89 elif DesType == 'SOCN': 102 elif destype == "All":
90 DS = protein.GetSOCN() 103 DS1 = protein.GetAAComp()
91 df = pd.DataFrame(DS, index=[0]) 104 DS2 = protein.GetDPComp()
92 elif DesType == 'TPComp': 105 DS3 = protein.GetTPComp()
93 DS = protein.GetTPComp() 106 DS4 = protein.GetMoreauBrotoAuto()
94 df = pd.DataFrame(DS, index=[0]) 107 DS5 = protein.GetMoranAuto()
95 elif DesType == 'All': 108 DS6 = protein.GetGearyAuto()
96 DS_1 = protein.GetPAAC(lamda=int(Lamda), weight=float(Weight)) 109 DS7 = protein.GetCTD()
97 DS_2 = protein.GetAPAAC(lamda=int(Lamda), weight=float(Weight)) 110 DS8 = protein.GetPAAC(lamda=int(lamda), weight=float(weight))
98 DS_3 = protein.GetCTD() 111 DS9 = protein.GetAPAAC(lamda=int(lamda), weight=float(weight))
99 DS_4 = protein.GetDPComp() 112 DS10 = protein.GetSOCN(maxlag=int(maxlag))
100 DS_5 = protein.GetGearyAuto() 113 DS11 = protein.GetQSO(maxlag=int(maxlag), weight=float(weight))
101 DS_6 = protein.GetMoranAuto() 114 DS12 = protein.GetTriad()
102 DS_7 = protein.GetMoreauBrotoAuto()
103 DS_8 = protein.GetQSO()
104 DS_9 = protein.GetSOCN()
105 DS_10 = protein.GetTPComp()
106 115
107 DS = {} 116 DS = {}
108 117
109 for D in (DS_1, DS_2, DS_3, DS_4, DS_5, DS_6, DS_7, DS_8, DS_9, DS_10): 118 for D in (DS1,DS2,DS3,DS4,DS5,DS6,DS7,DS8,DS9,DS10,DS11,DS12):
119 print(D)
110 DS.update(D) 120 DS.update(D)
111
112 df = pd.DataFrame(DS, index=[0]) 121 df = pd.DataFrame(DS, index=[0])
113 122
114 if DesType == 'BinaryDescriptor': 123 if destype == 'BinaryDescriptor':
115 out_df = BinaryDescriptor(list_pep_name) 124 out_df = BinaryDescriptor(list_pep_name)
116 else: 125 else:
117 out_df = pd.concat([out_df, df], axis=0) 126 out_df = pd.concat([out_df, df], axis=0)
118 127
119 128
120 out_df.to_csv(Out_file, index=False, sep='\t') 129 out_df.to_csv(out_file, index=False, sep='\t')
121 130
122 131
123 if __name__=="__main__": 132 if __name__=="__main__":
124 133
125 134
139 148
140 parser.add_argument("-w", "--Weight", 149 parser.add_argument("-w", "--Weight",
141 required=False, 150 required=False,
142 default=0.5, 151 default=0.5,
143 help="pep file") 152 help="pep file")
144 153
154 parser.add_argument("-m", "--MaxLag",
155 required=False,
156 default=10,
157 help="pep file")
158
145 parser.add_argument("-t", "--DesType", 159 parser.add_argument("-t", "--DesType",
146 required=True, 160 required=True,
147 default=None, 161 default=None,
148 help="out put file name for str Descriptors") 162 help="out put file name for str Descriptors")
149 163
151 required=False, 165 required=False,
152 default="Out.tsv", 166 default="Out.tsv",
153 help="Path to target tsv file") 167 help="Path to target tsv file")
154 168
155 args = parser.parse_args() 169 args = parser.parse_args()
156 Decriptor_generator(args.InFile, args.Lamda, args.Weight, args.DesType, args.Out_file) 170 Decriptor_generator(args.InFile, args.Lamda, args.Weight, args.MaxLag, args.DesType, args.Out_file)
157 171
158 172