diff options
| -rw-r--r-- | wqflask/wqflask/correlation/show_corr_results.py | 125 |
1 files changed, 74 insertions, 51 deletions
diff --git a/wqflask/wqflask/correlation/show_corr_results.py b/wqflask/wqflask/correlation/show_corr_results.py index 11eca936..4a9aea73 100644 --- a/wqflask/wqflask/correlation/show_corr_results.py +++ b/wqflask/wqflask/correlation/show_corr_results.py @@ -52,6 +52,8 @@ import utility.webqtlUtil #this is for parallel computing only. from wqflask.correlation import correlationFunction from utility.benchmark import Bench +from MySQLdb import escape_string as escape + from pprint import pformat as pf METHOD_SAMPLE_PEARSON = "1" @@ -101,13 +103,14 @@ class CorrelationResults(object): self.sample_data = {} self.corr_method = start_vars['corr_sample_method'] + self.return_number = 500 #The two if statements below append samples to the sample list based upon whether the user #rselected Primary Samples Only, Other Samples Only, or All Samples primary_samples = (self.dataset.group.parlist + - self.dataset.group.f1list + - self.dataset.group.samplelist) + self.dataset.group.f1list + + self.dataset.group.samplelist) #If either BXD/whatever Only or All Samples, append all of that group's samplelist if corr_samples_group != 'samples_other': @@ -153,7 +156,7 @@ class CorrelationResults(object): #self.correlation_data_slice = collections.OrderedDict() - for trait_counter, trait in enumerate(self.correlation_data.keys()[:300]): + for trait_counter, trait in enumerate(self.correlation_data.keys()[:self.return_number]): trait_object = GeneralTrait(dataset=self.dataset, name=trait, get_qtl_info=True) trait_object.sample_r = self.correlation_data[trait][0] trait_object.sample_p = self.correlation_data[trait][1] @@ -203,13 +206,6 @@ class CorrelationResults(object): # mb = trait_object.mb # ) - - - #trait_list = self.getTissueCorrelationByList( primary_trait_symbol = self.this_trait.symbol, - # corr_results = self.correlation_results, - # TissueProbeSetFreezeId = 1, - # method=1) - #XZ, 09/18/2008: get all information about the user selected database. #target_db_name = fd.corr_dataset #self.target_db_name = start_vars['corr_dataset'] @@ -531,9 +527,13 @@ class CorrelationResults(object): #XZ, 09/23/2008: In tissue correlation tables, there is no record of GeneId1 == GeneId2 #XZ, 09/24/2008: Note that the correlation value can be negative. - def getTempTissueCorrTable(self, primaryTraitSymbol="", TissueProbeSetFreezeId=0, method="", returnNumber=0): + def get_temp_tissue_corr_table(self, + tissue_probesetfreeze_id=0, + method="", + return_number=0): + - def cmpTissCorrAbsoluteValue(A, B): + def cmp_tisscorr_absolute_value(A, B): try: if abs(A[1]) < abs(B[1]): return 1 elif abs(A[1]) == abs(B[1]): @@ -542,26 +542,28 @@ class CorrelationResults(object): except: return 0 - symbolCorrDict, symbolPvalueDict = self.calculateCorrOfAllTissueTrait(primaryTraitSymbol=primaryTraitSymbol, TissueProbeSetFreezeId=TISSUE_MOUSE_DB, method=method) + symbol_corr_dict, symbol_pvalue_dict = self.calculate_corr_all_tissue_trait( + tissue_probesetfreeze_id=TISSUE_MOUSE_DB, + method=method) - symbolCorrList = symbolCorrDict.items() + symbol_corr_list = symbol_corr_dict.items() - symbolCorrList.sort(cmpTissCorrAbsoluteValue) - symbolCorrList = symbolCorrList[0 : 2*returnNumber] + symbol_corr_list.sort(cmp_tisscorr_absolute_value) + symbol_corr_list = symbol_corr_list[0 : 2*return_number] - tmpTableName = webqtlUtil.genRandStr(prefix="TOPTISSUE") + tmp_table_name = webqtlUtil.genRandStr(prefix="TOPTISSUE") - q1 = 'CREATE TEMPORARY TABLE %s (Symbol varchar(100) PRIMARY KEY, Correlation float, PValue float)' % tmpTableName + q1 = 'CREATE TEMPORARY TABLE %s (Symbol varchar(100) PRIMARY KEY, Correlation float, PValue float)' % tmp_table_name self.cursor.execute(q1) - for one_pair in symbolCorrList: + for one_pair in symbol_corr_list: one_symbol = one_pair[0] one_corr = one_pair[1] - one_p_value = symbolPvalueDict[one_symbol] + one_p_value = symbol_pvalue_dict[one_symbol] self.cursor.execute( "INSERT INTO %s (Symbol, Correlation, PValue) VALUES ('%s',%f,%f)" % (tmpTableName, one_symbol, float(one_corr), float(one_p_value)) ) - return tmpTableName + return tmp_table_name #XZ, 01/09/2009: This function was created by David Crowell. Xiaodong cleaned up and modified it. @@ -591,37 +593,47 @@ class CorrelationResults(object): return litCorrDict + def fetch_tissue_correlations(self, method=""): + """Comments Possibly Out of Date!!!!! + + + Uses getTempTissueCorrTable to generate table of tissue correlations. + + This function then gathers that data and pairs it with the TraitID string. + Takes as its arguments a formdata instance, and a database instance. + Returns a dictionary of 'TraitID':(tissueCorr, tissuePValue) + for the requested correlation + + """ - #XZ, 01/09/2009: Xiaodong created this function. - def fetchTissueCorrelations(self, db, primaryTraitSymbol="", TissueProbeSetFreezeId=0, method="", returnNumber = 0): - """Uses getTempTissueCorrTable to generate table of tissue correlations. This function then gathers that data and - pairs it with the TraitID string. Takes as its arguments a formdata instance, and a database instance. - Returns a dictionary of 'TraitID':(tissueCorr, tissuePValue) for the requested correlation""" - - - tempTable = self.getTempTissueCorrTable(primaryTraitSymbol=primaryTraitSymbol, TissueProbeSetFreezeId=TISSUE_MOUSE_DB, method=method, returnNumber=returnNumber) + # table name string + temp_table = self.get_temp_tissue_corr_table(tissue_probesetfreeze_id=TISSUE_MOUSE_DB, + method=method) - query = "SELECT ProbeSet.Name, %s.Correlation, %s.PValue" % (tempTable, tempTable) - query += ' FROM (ProbeSet, ProbeSetXRef, ProbeSetFreeze)' - query += ' LEFT JOIN %s ON %s.Symbol=ProbeSet.Symbol ' % (tempTable,tempTable) - query += "WHERE ProbeSetFreeze.Name = '%s' and ProbeSetFreeze.Id=ProbeSetXRef.ProbeSetFreezeId and ProbeSet.Id = ProbeSetXRef.ProbeSetId and ProbeSet.Symbol IS NOT NULL AND %s.Correlation IS NOT NULL" % (db.name, tempTable) + query = """SELECT ProbeSet.Name, {}.Correlation, {}.PValue + FROM (ProbeSet, ProbeSetXRef, ProbeSetFreeze) + LEFT JOIN {} ON {}.Symbol=ProbeSet.Symbol + WHERE ProbeSetFreeze.Name = '{}' + and ProbeSetFreeze.Id=ProbeSetXRef.ProbeSetFreezeId + and ProbeSet.Id = ProbeSetXRef.ProbeSetId + and ProbeSet.Symbol IS NOT NULL + and {}.Correlation IS NOT NULL""".format(dataset.mescape( + temp_table, temp_table, temp_table, temp_table, + self.dataset.name, temp_table)) - self.cursor.execute(query) - results = self.cursor.fetchall() + results = g.db.execute(query).fetchall() - tissueCorrDict = {} + tissue_corr_dict = {} for entry in results: - traitName, tissueCorr, tissuePValue = entry - tissueCorrDict[traitName] = (tissueCorr, tissuePValue) - - self.cursor.execute('DROP TEMPORARY TABLE %s' % tempTable) + trait_name, tissue_corr, tissue_pvalue = entry + tissue_corr_dict[trait_name] = (tissue_corr, tissue_pvalue) - return tissueCorrDict + g.db.execute('DROP TEMPORARY TABLE {}'.format(escape(temp_table)) + return tissue_corr_dict - #XZ, 01/13/2008 def getLiteratureCorrelationByList(self, input_trait_mouse_geneid=None, species=None, traitList=None): tmpTableName = webqtlUtil.genRandStr(prefix="LITERATURE") @@ -672,7 +684,7 @@ class CorrelationResults(object): use_tissue_corr = False if self.method in TISSUE_METHODS: - tissue_corrs = self.fetchTissueCorrelations(db=self.db, primaryTraitSymbol=self.trait_symbol, TissueProbeSetFreezeId=TISSUE_MOUSE_DB, method=self.method, returnNumber = self.returnNumber) + tissue_corrs = self.fetch_tissue_correlations(method=self.method, return_number = self.return_number) use_tissue_corr = True DatabaseFileName = self.getFileName( target_db_name=self.target_db_name ) @@ -897,20 +909,31 @@ class CorrelationResults(object): return trait_list """ - def calculateCorrOfAllTissueTrait(self, primaryTraitSymbol=None, TissueProbeSetFreezeId=None, method=None): + def calculate_corr_all_tissue_trait(self, tissue_probesetfreeze_id=None, method=None): - symbolCorrDict = {} - symbolPvalueDict = {} + symbol_corr_dict = {} + symbol_pvalue_dict = {} - primaryTraitSymbolValueDict = correlationFunction.getGeneSymbolTissueValueDictForTrait(cursor=self.cursor, GeneNameLst=[primaryTraitSymbol], TissueProbeSetFreezeId=TISSUE_MOUSE_DB) - primaryTraitValue = primaryTraitSymbolValueDict.values()[0] + primary_trait_symbol_value_dict = correlation_function. + get_genesymbol_tissue_value_dict_trait(cursor=self.cursor, + GeneNameLst=[primaryTraitSymbol], + TissueProbeSetFreezeId=TISSUE_MOUSE_DB) + primary_trait_value = primary_trait_symbol_value_dict.values()[0] - SymbolValueDict = correlationFunction.getGeneSymbolTissueValueDictForTrait(cursor=self.cursor, GeneNameLst=[], TissueProbeSetFreezeId=TISSUE_MOUSE_DB) + symbol_value_dict = correlation_function.get_genesymbol_tissue_value_dict_trait( + cursor=self.cursor, + gene_name_list=[], + tissue_probeSetfreeze_id=TISSUE_MOUSE_DB) if method in ["2","5"]: - symbolCorrDict, symbolPvalueDict = correlationFunction.batchCalTissueCorr(primaryTraitValue,SymbolValueDict,method='spearman') + symbol_corr_dict, symbol_pvalue_dict = correlation_function.batch_cal_tissue_corr( + primaryTraitValue, + SymbolValueDict, + method='spearman') else: - symbolCorrDict, symbolPvalueDict = correlationFunction.batchCalTissueCorr(primaryTraitValue,SymbolValueDict) + symbol_corr_dict, symbol_pvalue_dict = correlation_function.batch_cal_tissue_corr( + primaryTraitValue, + SymbolValueDict) return (symbolCorrDict, symbolPvalueDict) |