diff get_proteins.py @ 0:e4b3fc88efe0 draft

Uploaded
author pedro_araujo
date Wed, 27 Jan 2021 13:50:11 +0000
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/get_proteins.py	Wed Jan 27 13:50:11 2021 +0000
@@ -0,0 +1,269 @@
+
+class PhageBacteriaInformation:
+
+	def __init__(self, dataset=None):
+		"""
+		Imports a dataset from NCBI Virus, where the columns are Phage ID, Phage Name, Bacteria Name, Bacteria ID.
+		If a phage entry does not have a bacteria associated, it is deleted
+		:param dataset:
+		"""
+		import pandas as pd
+		import ast
+		self.phagesProteins = {}
+		# self.phagesDNA = {}
+		self.bactProteins = {}
+		# self.bactDNA = {}
+		self.data = pd.read_csv('files/'+dataset, header=0, index_col=0)
+		self.data = self.data.dropna(how='any')
+		self.data = self.data[self.data['Host_ID'] != '[]']
+		index_remove = []
+		for i in range(len(self.data)):
+			temp = self.data['Host'][i].split(' ')
+			if len(temp) <= 1:
+				index_remove.append(i)
+		self.data = self.data.drop(self.data.index[index_remove])
+		self.all_bact = []
+		for i in self.data.index:
+			for bact in ast.literal_eval(self.data.loc[i, 'Host_ID']):
+				if bact[:-2] not in self.all_bact:
+					self.all_bact.append(bact[:-2])
+		self.data.to_csv('files/Filtered_Phage_Bacteria.csv')
+
+	def addFeatures(self):
+		"""
+		For each phage in the data, it saves its DNA sequence and all proteins, as provided by NCBI. It saves them into two variables
+		Each bacteria associated with the phage is also searched for its DNA and proteins sequences.
+		:return:
+		"""
+		from Bio import Entrez
+		from Bio import SeqIO
+		import json
+		import ast
+		Entrez.email = 'pedro_araujo97@hotmail.com'
+		print('Working...')
+		for phageID in self.data.index:
+			with Entrez.efetch(db="nucleotide", rettype="gb", retmode="text", id=phageID) as handle:
+				genomePhage = SeqIO.read(handle, "gb")
+			protsPhage = {}
+			for feat in genomePhage.features:
+				if feat.type == 'CDS':
+					try: protsPhage[feat.qualifiers['protein_id'][0]] = [feat.qualifiers['product'][0], feat.qualifiers['translation'][0]]
+					except: pass
+			self.phagesProteins[phageID] = protsPhage
+
+		for bact in self.all_bact:
+			protsBac = {}
+			with Entrez.efetch(db="nucleotide", rettype="gb", retmode="text", id=bact) as handle:
+				genomeBac = SeqIO.read(handle, "gb")
+			for feat in genomeBac.features:
+				if feat.type == 'CDS':
+					try: protsBac[feat.qualifiers['protein_id'][0]] = [feat.qualifiers['product'][0], feat.qualifiers['translation'][0]]
+					except: pass
+			self.bactProteins[bact] = protsBac
+
+		with open('files/phagesProteins.json', 'w') as f:
+			json.dump(self.phagesProteins, f)
+		self.__createFasta(self.phagesProteins, 'phagesProteins')
+		with open('files/bactProteins.json', 'w') as f:
+			json.dump(self.bactProteins, f)
+		self.__createFasta(self.bactProteins, 'bactProteins')
+		print('Done')
+
+	def addBacteriaFeatures(self):
+		"""
+		For each unique bacteria present in the dataset, the DNA and protein sequences are saved in two variables.
+		:return:
+		"""
+		from Bio import Entrez
+		from Bio import SeqIO
+		import json
+		import ast
+		Entrez.email = 'pedro_araujo97@hotmail.com'
+		print('Working...')
+		for bact in self.all_bact:
+			if bact not in self.bactProteins.keys():
+				protsBac = {}
+				try:
+					with Entrez.efetch(db="nucleotide", rettype="gb", retmode="text", id=bact) as handle:
+						genomeBac = SeqIO.read(handle, "gb")
+					for feat in genomeBac.features:
+						if feat.type == 'CDS':
+							try: protsBac[feat.qualifiers['protein_id'][0]] = [feat.qualifiers['product'][0], feat.qualifiers['translation'][0]]
+							except: pass
+					if len(genomeBac.features) <= 5:
+						with Entrez.efetch(db="nucleotide", rettype="gbwithparts", retmode="text", id=bact) as handle:
+							genomeBac = handle.readlines()
+						for i in range(len(genomeBac)):
+							if ' CDS ' in genomeBac[i]:
+								j = i
+								protDone = False
+								while j < len(genomeBac):
+									if protDone:
+										break
+									if '/product=' in genomeBac[j]:
+										product = genomeBac[j].strip()[10:]
+										j += 1
+									elif '_id=' in genomeBac[j]:
+										protKey = genomeBac[j].strip()[13:-1]
+										j += 1
+									elif '/translation=' in genomeBac[j]:
+										protSeq = genomeBac[j].strip()[14:]
+										j += 1
+										for k in range(j, len(genomeBac)):
+											if genomeBac[k].islower():
+												j = k
+												protDone = True
+												break
+											else:
+												protSeq += genomeBac[k].strip()
+									else:
+										j += 1
+								protsBac[protKey] = [product, protSeq[:protSeq.find('"')]]
+					self.bactProteins[bact] = protsBac
+				except:
+					print(bact + ' failed')
+		with open('files/bactProteins.json', 'w') as f:
+			json.dump(self.bactProteins, f)
+		self.__createFasta(self.bactProteins, 'bactProteins')
+		print('Done')
+
+	def add_individual_bacteria(self):
+		"""
+		For each unique bacteria present in the dataset, the DNA and protein sequences are saved in two variables.
+		:return:
+		"""
+		from Bio import Entrez
+		from Bio import SeqIO
+		import json
+		from pathlib import Path
+		Entrez.email = 'pedro_araujo97@hotmail.com'
+		print('Working...')
+		for bact in self.all_bact:
+			my_file = Path('files/bacteria/' + bact + ".json")
+			if not my_file.is_file():
+				protsBac = {}
+				try:
+					with Entrez.efetch(db="nucleotide", rettype="gb", retmode="text", id=bact) as handle:
+						genomeBac = SeqIO.read(handle, "gb")
+					for feat in genomeBac.features:
+						if feat.type == 'CDS':
+							try: protsBac[feat.qualifiers['protein_id'][0]] = [feat.qualifiers['product'][0], feat.qualifiers['translation'][0]]
+							except: pass
+					if len(genomeBac.features) <= 5:
+						with Entrez.efetch(db="nucleotide", rettype="gbwithparts", retmode="text", id=bact) as handle:
+							genomeBac = handle.readlines()
+						for i in range(len(genomeBac)):
+							if ' CDS ' in genomeBac[i]:
+								j = i
+								protDone = False
+								while j < len(genomeBac):
+									if protDone:
+										break
+									if '/product=' in genomeBac[j]:
+										product = genomeBac[j].strip()[10:]
+										j += 1
+									elif '_id=' in genomeBac[j]:
+										protKey = genomeBac[j].strip()[13:-1]
+										j += 1
+									elif '/translation=' in genomeBac[j]:
+										protSeq = genomeBac[j].strip()[14:]
+										j += 1
+										for k in range(j, len(genomeBac)):
+											if genomeBac[k].islower():
+												j = k
+												protDone = True
+												break
+											else:
+												protSeq += genomeBac[k].strip()
+									else:
+										j += 1
+								protsBac[protKey] = [product, protSeq[:protSeq.find('"')]]
+					with open('files/bacteria/' + bact + '.json', 'w') as f:
+						json.dump(protsBac, f)
+				except:
+					print(bact + ' failed')
+		# with open('files/bactProteins.json', 'w') as f:
+		# 	json.dump(self.bactProteins, f)
+		# self.__createFasta(self.bactProteins, 'bactProteins')
+		print('Done')
+
+	def importData(self):
+		"""
+		Imports the previously saved DNA and protein sequences. This needs to be improved so the user can specify which data to import.
+		:return:
+		"""
+		import json
+		with open('files/phagesProteins.json', encoding='utf-8') as F:
+			self.phagesProteins = json.loads(F.read())
+		# with open('files/bactProteins.json', encoding='utf-8') as F:
+		# 	self.bactProteins = json.loads(F.read())
+
+	def addPhageProt(self):
+		"""
+		For each unique phage present in the dataset, the DNA and protein sequences are saved in two variables.
+		:return:
+		"""
+		from Bio import Entrez
+		from Bio import SeqIO
+		import json
+		Entrez.email = 'pedro_araujo97@hotmail.com'
+		print('Working...')
+
+		for phageID in self.data.index:
+			with Entrez.efetch(db="nucleotide", rettype="gb", retmode="text", id=phageID) as handle:
+				genomePhage = SeqIO.read(handle, "gb")
+			protsPhage = {}
+			for feat in genomePhage.features:
+				if feat.type == 'CDS':
+					try: protsPhage[feat.qualifiers['protein_id'][0]] = [feat.qualifiers['product'][0], feat.qualifiers['translation'][0]]
+					except: pass
+			self.phagesProteins[phageID] = protsPhage
+
+		with open('files/phagesProteins.json', 'w') as f:
+			json.dump(self.phagesProteins, f)
+		self.__createFasta(self.phagesProteins, 'phagesProteins')
+		return self.phagesProteins
+
+	def add_missing_phage(self):
+		"""
+		For each unique phage present in the dataset, the DNA and protein sequences are saved in two variables.
+		:return:
+		"""
+		from Bio import Entrez
+		from Bio import SeqIO
+		import json
+		Entrez.email = 'pedro_araujo97@hotmail.com'
+		print('Working...')
+
+		for phageID in self.data.index:
+			if phageID not in self.phagesProteins.keys():
+				print(phageID)
+				with Entrez.efetch(db="nucleotide", rettype="gb", retmode="text", id=phageID) as handle:
+					genomePhage = SeqIO.read(handle, "gb")
+				protsPhage = {}
+				for feat in genomePhage.features:
+					if feat.type == 'CDS':
+						try: protsPhage[feat.qualifiers['protein_id'][0]] = [feat.qualifiers['product'][0], feat.qualifiers['translation'][0]]
+						except: pass
+				self.phagesProteins[phageID] = protsPhage
+		with open('files/phagesProteins.json', 'w') as f:
+			json.dump(self.phagesProteins, f)
+		self.__createFasta(self.phagesProteins, 'phagesProteins')
+		return self.phagesProteins
+
+	def __createFasta(self, var, name):
+		with open('files/' + name + '.fasta', 'w') as F:
+			for spec in var:
+				try:
+					for prot in var[spec]:
+						F.write('>' + prot + '-' + spec + '\n' + var[spec][prot][1] + '\n')
+				except:
+					F.write('>' + spec + '\n' + var[spec] + '\n')
+
+
+if __name__ == '__main__':
+	test = PhageBacteriaInformation('NCBI_Phage_Bacteria_Data.csv')
+	test.add_individual_bacteria()
+	test.addPhageProt()
+	test.add_missing_phage()
+	test.importData()