def run(self):
if self.debug:
import pdb
pdb.set_trace()
db = Stock_250kDB(drivername=self.drivername, username=self.db_user,
password=self.db_passwd, hostname=self.hostname, database=self.dbname, schema=self.schema)
db.setup()
session = db.session
results_method_id_info = self.getResultsMethodIDInfo(db, self.call_method_id_ls, self.min_distance, self.get_closest, self.min_MAF)
results_method_id2gene_set = self.getResultsMethodID2GeneSet(db, results_method_id_info, self.results_directory, self.max_rank)
rdata = self.getDataMatrix(results_method_id2gene_set, results_method_id_info)
header = ['', ''] + results_method_id_info.results_method_id_label_ls
strain_acc_list = results_method_id_info.results_method_id_label_ls
category_list = results_method_id_info.results_method_id_ls
if SNPData.isDataMatrixEmpty(rdata.data_matrix):
sys.stderr.write("Nothing fetched from database.\n")
sys.exit(3)
if self.output_fname:
write_data_matrix(rdata.data_matrix, self.output_fname, header, strain_acc_list, category_list)
if self.fig_fname:
font = get_font(self.font_path, font_size=self.font_size) #2008-08-01
value2color_func = lambda x: Value2Color.value2HSLcolor(x, rdata.min_value, rdata.max_value)
im_legend = drawContinousLegend(rdata.min_value, rdata.max_value, self.no_of_ticks, value2color_func, font)
#im.save('%s_legend.png'%self.fig_fname_prefix)
im = drawMatrix(rdata.data_matrix, value2color_func, strain_acc_list,\
strain_acc_list, with_grid=1, font=font)
im = combineTwoImages(im, im_legend, font=font)
im.save(self.fig_fname)
def run(self):
if self.debug:
import pdb
pdb.set_trace()
db = StockDB.StockDB(drivername=self.drivername, username=self.db_user,
password=self.db_passwd, hostname=self.hostname, database=self.dbname, schema=self.schema)
db.setup(create_tables=False)
session = db.session
order_by_sentence = " order by c.longitude, c.latitude, e.longitude, e.latitude, e.nativename " #how to order strains.
if self.QC_method_id ==4:
sql_table_str = "from %s e, %s s, %s a, %s c"%(StockDB.Ecotype.table.name, StockDB.Site.table.name, StockDB.Address.table.name,\
StockDB.Country.table.name)
common_where_condition = "where e.siteid=s.id and s.addressid=a.id and a.countryid=c.id %s " + order_by_sentence
strain_where_condition = common_where_condition%(" and e.id=st.ecotypeid")
strain_id_info_query = "select distinct st.id as strainid, e.id as ecotypeid, e.nativename, s.name as sitename, c.abbr %s, %s st %s"%(sql_table_str, StockDB.Strain.table.name, strain_where_condition)
else:
sql_table_str = "from %s q, %s e, %s s, %s a, %s c"%(StockDB.QCCrossMatch.table.name, StockDB.Ecotype.table.name, StockDB.Site.table.name, StockDB.Address.table.name,\
StockDB.Country.table.name)
common_where_condition = "where e.siteid=s.id and s.addressid=a.id and a.countryid=c.id %s"+ " and q.qc_method_id=%s and q.no_of_non_NA_pairs>=%s and q.mismatch_rate<=%s "%\
(self.QC_method_id, self.min_no_of_non_NAs, self.max_mismatch_rate) + order_by_sentence
strain_where_condition = common_where_condition%(" and e.id=st.ecotypeid and st.id=q.strainid")
strain_id_info_query = "select distinct q.strainid, e.id as ecotypeid, e.nativename, s.name as sitename, c.abbr %s, %s st %s"%(sql_table_str, StockDB.Strain.table.name, strain_where_condition)
if self.how_to_group_strains==2 or self.how_to_group_strains==3:
plate_info = self.alignStrainsAccordingToSeqPlate(db)
id_set_data = PassingData()
id_set_data.strain_id_set = None
id_set_data.target_id_set = None
elif self.input_fname:
id_set_data = self.getStrainidTargetidFromFile(db, self.QC_method_id, self.input_fname, self.max_mismatch_rate, self.min_no_of_non_NAs)
else:
id_set_data = PassingData()
id_set_data.strain_id_set = None
id_set_data.target_id_set = None
if self.how_to_group_strains==2 or self.how_to_group_strains==3:
strain_id_info = self.getStrainInfoGivenPlateInfo(db, plate_info, strain_id_info_query, strain_id_set=None)
else:
strain_id_info = self.getStrainIDInfo(db, strain_id_info_query, id_set_data.strain_id_set)
if self.QC_method_id==4:
if self.how_to_group_strains==3:
#2008-09-15 column strain id is in country, strain-longitude order
target_id_info = self.getStrainIDInfo(db, strain_id_info_query, id_set_data.strain_id_set)
else:
target_id_info = strain_id_info
else:
target_where_condition = common_where_condition%(" and e.id=q.target_id")
target_id_info_query = "select distinct e.id as strainid, e.id as ecotypeid, e.nativename, s.name as sitename, c.abbr %s %s"%(sql_table_str, target_where_condition)
target_id_info = self.getStrainIDInfo(db, target_id_info_query)
if self.input_fname:
rdata = self.get_data_matrixFromFile(db, strain_id_info, target_id_info, self.QC_method_id, self.input_fname, self.max_mismatch_rate, self.min_no_of_non_NAs)
else:
rdata = self.get_data_matrix(db, strain_id_info, target_id_info, self.QC_method_id, self.max_mismatch_rate, self.min_no_of_non_NAs)
rdata.data_matrix = self.markDataMatrixBoundary(rdata.data_matrix, strain_id_info, target_id_info)
header = ['strain info', ''] + target_id_info.strain_label_ls
strain_acc_list = strain_id_info.strain_label_ls
category_list = [1]*len(strain_acc_list)
if SNPData.isDataMatrixEmpty(rdata.data_matrix):
sys.stderr.write("Nothing fetched from database.\n")
sys.exit(3)
if self.output_fname:
write_data_matrix(rdata.data_matrix, self.output_fname, header, strain_acc_list, category_list)
if self.fig_fname:
font = get_font(self.font_path, font_size=self.font_size) #2008-08-01
value2color_func = lambda x: Value2Color.value2HSLcolor(x, rdata.min_value, rdata.max_value)
im_legend = drawContinousLegend(rdata.min_value, rdata.max_value, self.no_of_ticks, value2color_func, font)
#im.save('%s_legend.png'%self.fig_fname_prefix)
im = drawMatrix(rdata.data_matrix, value2color_func, strain_id_info.strain_label_ls,\
target_id_info.strain_label_ls, with_grid=1, font=font)
im = combineTwoImages(im, im_legend, font=font)
im.save(self.fig_fname)
def run(self):
if self.debug:
import pdb
pdb.set_trace()
db = StockDB.StockDB(drivername=self.drivername,
username=self.db_user,
password=self.db_passwd,
hostname=self.hostname,
database=self.dbname,
schema=self.schema)
db.setup(create_tables=False)
session = db.session
order_by_sentence = " order by c.longitude, c.latitude, e.longitude, e.latitude, e.nativename " #how to order strains.
if self.QC_method_id == 4:
sql_table_str = "from %s e, %s s, %s a, %s c"%(StockDB.Ecotype.table.name, StockDB.Site.table.name, StockDB.Address.table.name,\
StockDB.Country.table.name)
common_where_condition = "where e.siteid=s.id and s.addressid=a.id and a.countryid=c.id %s " + order_by_sentence
strain_where_condition = common_where_condition % (
" and e.id=st.ecotypeid")
strain_id_info_query = "select distinct st.id as strainid, e.id as ecotypeid, e.nativename, s.name as sitename, c.abbr %s, %s st %s" % (
sql_table_str, StockDB.Strain.table.name,
strain_where_condition)
else:
sql_table_str = "from %s q, %s e, %s s, %s a, %s c"%(StockDB.QCCrossMatch.table.name, StockDB.Ecotype.table.name, StockDB.Site.table.name, StockDB.Address.table.name,\
StockDB.Country.table.name)
common_where_condition = "where e.siteid=s.id and s.addressid=a.id and a.countryid=c.id %s"+ " and q.qc_method_id=%s and q.no_of_non_NA_pairs>=%s and q.mismatch_rate<=%s "%\
(self.QC_method_id, self.min_no_of_non_NAs, self.max_mismatch_rate) + order_by_sentence
strain_where_condition = common_where_condition % (
" and e.id=st.ecotypeid and st.id=q.strainid")
strain_id_info_query = "select distinct q.strainid, e.id as ecotypeid, e.nativename, s.name as sitename, c.abbr %s, %s st %s" % (
sql_table_str, StockDB.Strain.table.name,
strain_where_condition)
if self.how_to_group_strains == 2 or self.how_to_group_strains == 3:
plate_info = self.alignStrainsAccordingToSeqPlate(db)
id_set_data = PassingData()
id_set_data.strain_id_set = None
id_set_data.target_id_set = None
elif self.input_fname:
id_set_data = self.getStrainidTargetidFromFile(
db, self.QC_method_id, self.input_fname,
self.max_mismatch_rate, self.min_no_of_non_NAs)
else:
id_set_data = PassingData()
id_set_data.strain_id_set = None
id_set_data.target_id_set = None
if self.how_to_group_strains == 2 or self.how_to_group_strains == 3:
strain_id_info = self.getStrainInfoGivenPlateInfo(
db, plate_info, strain_id_info_query, strain_id_set=None)
else:
strain_id_info = self.getStrainIDInfo(db, strain_id_info_query,
id_set_data.strain_id_set)
if self.QC_method_id == 4:
if self.how_to_group_strains == 3:
#2008-09-15 column strain id is in country, strain-longitude order
target_id_info = self.getStrainIDInfo(
db, strain_id_info_query, id_set_data.strain_id_set)
else:
target_id_info = strain_id_info
else:
target_where_condition = common_where_condition % (
" and e.id=q.target_id")
target_id_info_query = "select distinct e.id as strainid, e.id as ecotypeid, e.nativename, s.name as sitename, c.abbr %s %s" % (
sql_table_str, target_where_condition)
target_id_info = self.getStrainIDInfo(db, target_id_info_query)
if self.input_fname:
rdata = self.get_data_matrixFromFile(db, strain_id_info,
target_id_info,
self.QC_method_id,
self.input_fname,
self.max_mismatch_rate,
self.min_no_of_non_NAs)
else:
rdata = self.get_data_matrix(db, strain_id_info, target_id_info,
self.QC_method_id,
self.max_mismatch_rate,
self.min_no_of_non_NAs)
rdata.data_matrix = self.markDataMatrixBoundary(
rdata.data_matrix, strain_id_info, target_id_info)
header = ['strain info', ''] + target_id_info.strain_label_ls
strain_acc_list = strain_id_info.strain_label_ls
category_list = [1] * len(strain_acc_list)
if SNPData.isDataMatrixEmpty(rdata.data_matrix):
sys.stderr.write("Nothing fetched from database.\n")
sys.exit(3)
if self.output_fname:
write_data_matrix(rdata.data_matrix, self.output_fname, header,
strain_acc_list, category_list)
if self.fig_fname:
font = get_font(self.font_path,
font_size=self.font_size) #2008-08-01
value2color_func = lambda x: Value2Color.value2HSLcolor(
x, rdata.min_value, rdata.max_value)
im_legend = drawContinousLegend(rdata.min_value, rdata.max_value,
self.no_of_ticks, value2color_func,
font)
#im.save('%s_legend.png'%self.fig_fname_prefix)
im = drawMatrix(rdata.data_matrix, value2color_func, strain_id_info.strain_label_ls,\
target_id_info.strain_label_ls, with_grid=1, font=font)
im = combineTwoImages(im, im_legend, font=font)
im.save(self.fig_fname)
def run(self):
if self.debug:
import pdb
pdb.set_trace()
db = Stock_250kDB.Stock_250kDB(drivername=self.drivername, username=self.db_user,
password=self.db_passwd, hostname=self.hostname, database=self.dbname, schema=self.schema)
db.setup()
session = db.session
if self.test_result_type==1:
test_result_class_table = CandidateGeneRankSumTestResult.table.name
test_result_class_table = 'candidate_gene_rank_sum_test_result_2008_09_15'
elif self.test_result_type==2:
test_result_class_table = CandidateGeneTopSNPTest.table.name
elif self.test_result_type==3:
test_result_class_table = Stock_250kDB.CandidateGeneRankSumTestResultMethod.table.name
else:
sys.stderr.write(" test_result_type %s not supported.\n"%(self.test_result_type))
sys.exit(2)
#the condition for min_MAF is tricky because of the floating precision.
if self.test_result_type==1:
where_condition = "%s r, %s c, %s g where g.id=c.list_type_id and r.analysis_method_id is not null \
and c.results_id=r.id and c.get_closest=%s and c.min_distance=%s and abs(c.min_MAF-%s)<0.00001"\
%(ResultsMethod.table.name, test_result_class_table, GeneListType.table.name, self.get_closest, self.min_distance, self.min_MAF)
elif self.test_result_type==2:
where_condition = "%s r, %s rg, %s c, %s g where g.id=c.list_type_id and r.analysis_method_id is not null and r.id=rg.results_method_id \
and c.results_id=rg.id and c.get_closest=%s and c.min_distance=%s and abs(c.min_MAF-%s)<0.00001"\
%(ResultsMethod.table.name, ResultsByGene.table.name, test_result_class_table, GeneListType.table.name, self.get_closest, self.min_distance, self.min_MAF)
elif self.test_result_type==3:
where_condition = "%s r, %s c, %s g where g.id=c.list_type_id and r.analysis_method_id is not null \
and c.results_id=r.id and c.get_closest=%s and c.min_distance=%s and abs(c.min_MAF-%s)<0.00001"\
%(ResultsMethod.table.name, test_result_class_table, GeneListType.table.name, self.get_closest, self.min_distance, self.min_MAF)
if self.call_method_id_ls:
where_condition += " and r.call_method_id in (%s)"%self.call_method_id_ls
if self.analysis_method_id_ls:
where_condition += " and r.analysis_method_id in (%s)"%self.analysis_method_id_ls
if self.super_type_id:
where_condition += " and g.super_type_id=%s"%self.super_type_id
if self.test_type:
where_condition += " and c.test_type=%s"%self.test_type
if self.test_result_type==1:
pass
where_condition += " and c.max_pvalue_per_gene=%s"%(self.max_pvalue_per_gene)
elif self.test_result_type==2:
where_condition += " and c.no_of_top_snps=%s"%(self.no_of_top_snps)
list_type_id_ls = self.getListTypeInfo(db, where_condition)
analysis_method_id_ls = self.getAnalysisMethodInfo(db, where_condition)
list_type_analysis_method_info = self.orderListTypeAnalysisMethodID(list_type_id_ls, analysis_method_id_ls)
phenotype_info = self.getPhenotypeInfo(db, where_condition)
rdata = self.get_data_matrix(db, phenotype_info, list_type_analysis_method_info, where_condition)
rdata.data_matrix = self.markDataMatrixBoundary(rdata.data_matrix, phenotype_info, list_type_analysis_method_info)
header = ['list_type_analysis_method', ''] + phenotype_info.phenotype_method_label_ls
strain_acc_list = list_type_analysis_method_info.list_type_analysis_method_label_ls
category_list = list_type_analysis_method_info.list_type_id_analysis_method_id_ls
if SNPData.isDataMatrixEmpty(rdata.data_matrix):
sys.stderr.write("Nothing fetched from database.\n")
sys.exit(3)
if self.output_fname:
write_data_matrix(rdata.data_matrix, self.output_fname, header, strain_acc_list, category_list)
if self.fig_fname:
font = get_font(self.font_path, font_size=self.font_size) #2008-08-01
value2color_func = lambda x: Value2Color.value2HSLcolor(x, rdata.min_value, rdata.max_value)
im_legend = drawContinousLegend(rdata.min_value, rdata.max_value, self.no_of_ticks, value2color_func, font)
#im.save('%s_legend.png'%self.fig_fname_prefix)
im = drawMatrix(rdata.data_matrix, value2color_func, list_type_analysis_method_info.list_type_analysis_method_label_ls,\
phenotype_info.phenotype_method_label_ls, with_grid=1, font=font)
im = combineTwoImages(im, im_legend, font=font)
im.save(self.fig_fname)
def run(self):
if self.debug:
import pdb
pdb.set_trace()
db = Stock_250kDB.Stock_250kDB(drivername=self.drivername,
username=self.db_user,
password=self.db_passwd,
hostname=self.hostname,
database=self.dbname,
schema=self.schema)
db.setup()
session = db.session
if self.test_result_type == 1:
test_result_class_table = CandidateGeneRankSumTestResult.table.name
test_result_class_table = 'candidate_gene_rank_sum_test_result_2008_09_15'
elif self.test_result_type == 2:
test_result_class_table = CandidateGeneTopSNPTest.table.name
elif self.test_result_type == 3:
test_result_class_table = Stock_250kDB.CandidateGeneRankSumTestResultMethod.table.name
else:
sys.stderr.write(" test_result_type %s not supported.\n" %
(self.test_result_type))
sys.exit(2)
#the condition for min_MAF is tricky because of the floating precision.
if self.test_result_type == 1:
where_condition = "%s r, %s c, %s g where g.id=c.list_type_id and r.analysis_method_id is not null \
and c.results_id=r.id and c.get_closest=%s and c.min_distance=%s and abs(c.min_MAF-%s)<0.00001" \
%(ResultsMethod.table.name, test_result_class_table, GeneListType.table.name, self.get_closest, self.min_distance, self.min_MAF)
elif self.test_result_type == 2:
where_condition = "%s r, %s rg, %s c, %s g where g.id=c.list_type_id and r.analysis_method_id is not null and r.id=rg.results_method_id \
and c.results_id=rg.id and c.get_closest=%s and c.min_distance=%s and abs(c.min_MAF-%s)<0.00001" \
%(ResultsMethod.table.name, ResultsByGene.table.name, test_result_class_table, GeneListType.table.name, self.get_closest, self.min_distance, self.min_MAF)
elif self.test_result_type == 3:
where_condition = "%s r, %s c, %s g where g.id=c.list_type_id and r.analysis_method_id is not null \
and c.results_id=r.id and c.get_closest=%s and c.min_distance=%s and abs(c.min_MAF-%s)<0.00001" \
%(ResultsMethod.table.name, test_result_class_table, GeneListType.table.name, self.get_closest, self.min_distance, self.min_MAF)
if self.call_method_id_ls:
where_condition += " and r.call_method_id in (%s)" % self.call_method_id_ls
if self.analysis_method_id_ls:
where_condition += " and r.analysis_method_id in (%s)" % self.analysis_method_id_ls
if self.super_type_id:
where_condition += " and g.super_type_id=%s" % self.super_type_id
if self.test_type:
where_condition += " and c.test_type=%s" % self.test_type
if self.test_result_type == 1:
pass
where_condition += " and c.max_pvalue_per_gene=%s" % (
self.max_pvalue_per_gene)
elif self.test_result_type == 2:
where_condition += " and c.no_of_top_snps=%s" % (
self.no_of_top_snps)
list_type_id_ls = self.getListTypeInfo(db, where_condition)
analysis_method_id_ls = self.getAnalysisMethodInfo(db, where_condition)
list_type_analysis_method_info = self.orderListTypeAnalysisMethodID(
list_type_id_ls, analysis_method_id_ls)
phenotype_info = self.getPhenotypeInfo(db, where_condition)
rdata = self.get_data_matrix(db, phenotype_info,
list_type_analysis_method_info,
where_condition)
rdata.data_matrix = self.markDataMatrixBoundary(
rdata.data_matrix, phenotype_info, list_type_analysis_method_info)
header = ['list_type_analysis_method', ''
] + phenotype_info.phenotype_method_label_ls
strain_acc_list = list_type_analysis_method_info.list_type_analysis_method_label_ls
category_list = list_type_analysis_method_info.list_type_id_analysis_method_id_ls
if SNPData.isDataMatrixEmpty(rdata.data_matrix):
sys.stderr.write("Nothing fetched from database.\n")
sys.exit(3)
if self.output_fname:
write_data_matrix(rdata.data_matrix, self.output_fname, header,
strain_acc_list, category_list)
if self.fig_fname:
font = get_font(self.font_path,
font_size=self.font_size) #2008-08-01
value2color_func = lambda x: Value2Color.value2HSLcolor(
x, rdata.min_value, rdata.max_value)
im_legend = drawContinousLegend(rdata.min_value, rdata.max_value,
self.no_of_ticks, value2color_func,
font)
#im.save('%s_legend.png'%self.fig_fname_prefix)
im = drawMatrix(rdata.data_matrix, value2color_func, list_type_analysis_method_info.list_type_analysis_method_label_ls,\
phenotype_info.phenotype_method_label_ls, with_grid=1, font=font)
im = combineTwoImages(im, im_legend, font=font)
im.save(self.fig_fname)