def TCGA_heatmap_pancancer_exist_sigGenes_private(target_cancer, sigGene_addr,tcga_cancer_diff_sig_gene_addr, tcga_heatmap_pancancer_addr, tcga_pancancer_cluster_addr):
with open(sigGene_addr) as sigGene_f:
significant_genes = [x.strip() for x in sigGene_f.readlines()]
# significant_genes = anova_sig_result_df['gene'].tolist()
print(significant_genes)
# anova_sig_result_df.to_csv("/home/wch23/Project/LifeArc/TCGA/Result/TCGA_GTEx_original_{}_vs_other_{}_anova_sig_genes_pv_{}.tsv".format(target_cancer, gene_value_mode,p_value_threshold),sep='\t')
# with open ("/home/wch23/Project/LifeArc/SOX2/result/Sig.Genes/{}_anova_sig_genes.txt".format(target_cancer),'w') as sig_f:
# sig_f.write('\n'.join(significant_genes))
#####
# draw heatmap with sig genes
#####
tcga_cancer_diff_df = dh.load_obj("/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_diff_matrix_original")
cancer_id_addr = "/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/cancer_id_info.tsv"
cancer_id_df = pd.read_csv(cancer_id_addr, sep='\t')
tcga_cancer_diff_df = tcga_cancer_diff_df[tcga_cancer_diff_df['gene_symbol'] != 'None']
tcga_cancer_diff_df = tcga_cancer_diff_df.drop(columns=['ensembl_id', 'ensembl_gene'])
tcga_cancer_original_order = list(tcga_cancer_diff_df)
tcga_cancer_diff_sig_gene_df = tcga_cancer_diff_df[tcga_cancer_diff_df['gene_symbol'].isin(significant_genes)]
print(tcga_cancer_diff_sig_gene_df.head())
print(cancer_id_df.head())
cancer_id_df = cancer_id_df.set_index('fullcode')
tcga_cancer_diff_sig_gene_df = tcga_cancer_diff_sig_gene_df.set_index('gene_symbol')
tcga_cancer_diff_sig_gene_cancer_type_df = pd.concat([tcga_cancer_diff_sig_gene_df.T, cancer_id_df],axis=1, join='inner' )
tcga_cancer_diff_sig_gene_cancer_type_df = tcga_cancer_diff_sig_gene_cancer_type_df.T
print(tcga_cancer_diff_sig_gene_cancer_type_df)
cancer_type = tcga_cancer_diff_sig_gene_cancer_type_df.loc['Abbreviation']
print(cancer_type)
rgb_colors = sns.color_palette("hls", len(cancer_type.unique()))
cancer_type_color = dict(zip(cancer_type.unique(), rgb_colors))
print (cancer_type_color)
col_colors = cancer_type.map(cancer_type_color)
#
print(tcga_cancer_diff_sig_gene_df.info())
tcga_cancer_diff_sig_gene_df.to_csv(tcga_cancer_diff_sig_gene_addr,sep='\t')
def TCGA_heatmap_pancancer_exist_sigGenes(target_cancer, sigGene_addr):
# # organism = '9606'
# # string_node_addr = "/home/wch23/Project/LifeArc/General/data/STRING/{}.protein.links.v11.0.400.nodes.txt".format(
# # organism)
# # with open(string_node_addr) as string_node_f:
# # string_node = [x.strip() for x in string_node_f.readlines()]
#
# ####
# # get sig genes from anova reslut
# ##########
#
# # target_cancer = 'PAAD'
# # target_cancer = "LUSC"
# gene_value_mode = 'Diff'
# # anova_result_addr = "/home/wch23/Project/LifeArc/TCGA/Result/TCGA_GTEx_{}_vs_other_{}_anova_result_with_ensembl.tsv".format(target_cancer,gene_value_mode)
# anova_result_addr = "/home/wch23/Project/LifeArc/TCGA/Result/TCGA_GTEx_original_{}_vs_other_{}_anova_result.csv".format(target_cancer, gene_value_mode)
# anova_result_df = pd.read_csv(anova_result_addr)
#
# # p_value_threshold = 1.0E-150
#
# anova_sig_result_df = anova_result_df.loc[anova_result_df['pvalue']<p_value_threshold]
# print(anova_sig_result_df.shape)
# print(anova_sig_result_df.head())
with open(sigGene_addr) as sigGene_f:
significant_genes = [x.strip() for x in sigGene_f.readlines()]
# significant_genes = anova_sig_result_df['gene'].tolist()
print(significant_genes)
# anova_sig_result_df.to_csv("/home/wch23/Project/LifeArc/TCGA/Result/TCGA_GTEx_original_{}_vs_other_{}_anova_sig_genes_pv_{}.tsv".format(target_cancer, gene_value_mode,p_value_threshold),sep='\t')
# with open ("/home/wch23/Project/LifeArc/SOX2/result/Sig.Genes/{}_anova_sig_genes.txt".format(target_cancer),'w') as sig_f:
# sig_f.write('\n'.join(significant_genes))
#####
# draw heatmap with sig genes
#####
tcga_cancer_diff_df = dh.load_obj("/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_diff_matrix_original")
cancer_id_addr = "/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/cancer_id_info.tsv"
cancer_id_df = pd.read_csv(cancer_id_addr, sep='\t')
tcga_cancer_diff_df = tcga_cancer_diff_df[tcga_cancer_diff_df['gene_symbol'] != 'None']
tcga_cancer_diff_df = tcga_cancer_diff_df.drop(columns=['ensembl_id', 'ensembl_gene'])
tcga_cancer_original_order = list(tcga_cancer_diff_df)
tcga_cancer_diff_sig_gene_df = tcga_cancer_diff_df[tcga_cancer_diff_df['gene_symbol'].isin(significant_genes)]
print(tcga_cancer_diff_sig_gene_df.head())
print(cancer_id_df.head())
cancer_id_df = cancer_id_df.set_index('fullcode')
tcga_cancer_diff_sig_gene_df = tcga_cancer_diff_sig_gene_df.set_index('gene_symbol')
tcga_cancer_diff_sig_gene_cancer_type_df = pd.concat([tcga_cancer_diff_sig_gene_df.T, cancer_id_df],axis=1, join='inner' )
tcga_cancer_diff_sig_gene_cancer_type_df = tcga_cancer_diff_sig_gene_cancer_type_df.T
print(tcga_cancer_diff_sig_gene_cancer_type_df)
cancer_type = tcga_cancer_diff_sig_gene_cancer_type_df.loc['Abbreviation']
print(cancer_type)
rgb_colors = sns.color_palette("hls", len(cancer_type.unique()))
cancer_type_color = dict(zip(cancer_type.unique(), rgb_colors))
print (cancer_type_color)
col_colors = cancer_type.map(cancer_type_color)
#
print(tcga_cancer_diff_sig_gene_df.info())
tcga_cancer_diff_sig_gene_df.to_csv('/home/wch23/Project/LifeArc/TCGA/Result/heatmap/{}/TCGA_diff_heatmap_matrix.tsv'.format(target_cancer),sep='\t')
g = sns.clustermap(tcga_cancer_diff_sig_gene_df, metric="correlation", cmap="RdBu_r", robust=True, method="average",z_score=0,
col_colors=col_colors,xticklabels=False) # Average is best
# g = sns.clustermap(tcga_cancer_diff_sig_gene_df, metric="correlation", cmap="RdBu_r", robust=True, method="average",
# col_colors=col_colors, xticklabels=False) # Average is best
g.savefig('/home/wch23/Project/LifeArc/TCGA/Result/heatmap/{}/TCGA_diff_heatmap_single_robust_colColor_norm.png'.format(target_cancer))
print(g.dendrogram_col.reordered_ind)
print(g.dendrogram_row.reordered_ind)
clustred_col = g.dendrogram_col.reordered_ind
cancer_type_df = tcga_cancer_diff_sig_gene_cancer_type_df.T[['Abbreviation']]
cancer_type_df.index.name = 'fullcode'
cancer_type_df = cancer_type_df.reset_index()
print(cancer_type_df)
clusterd_cancer_type_df = cancer_type_df.reindex(clustred_col)
print(clusterd_cancer_type_df)
clusterd_cancer_type_df.to_csv('/home/wch23/Project/LifeArc/TCGA/Result/heatmap/{}/TCGA_diff_heatmap_tcga_clustred_result.csv'.format(target_cancer))
def main_for_nonType():
tcga_cancer_diff_df = dh.load_obj(
"/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_diff_matrix_original"
)
with open(
'/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_types_has_normal.tsv'
) as cancer_type_f:
tcga_cancer_types = [x.strip() for x in cancer_type_f.readlines()]
cancer_id_addr = "/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/cancer_id_info.tsv"
cancer_id_df = pd.read_csv(cancer_id_addr, sep='\t')
# print(cancer_id_df)
# target_cancer_list=["PAAD"] # pancreas, ITGB5
# cancer_id_df['Case_Ctrl'] = ['Case' if x in target_cancer_list else 'Ctrl' for x in cancer_id_df['Abbreviation']]
# print(cancer_id_df)
#
tcga_cancer_diff_df = tcga_cancer_diff_df[
tcga_cancer_diff_df['gene_symbol'] != 'None']
tcga_cancer_diff_df = tcga_cancer_diff_df.drop(
columns=['ensembl_id', 'ensembl_gene'])
# gene_symbol_list = ["MIR3648-2","AL513534.2","MIR3648-1","MIR6753","AC099677.4","EXOSC8","AC127024.5","FAT3","SCARB2","PABPC4L","RUSC1-AS1","GDNF-AS1","IVD","SLC39A2","CBWD6","DUSP12","SMC6","AL031587.5","SBNO2"]
gene_symbol_list = ["ERBB2"]
cancer_type = 'All'
cancer_id_df = cancer_id_df.set_index('fullcode')
for gene_symbol in gene_symbol_list:
aGene_TCGA_df = tcga_cancer_diff_df.loc[
tcga_cancer_diff_df['gene_symbol'] == gene_symbol]
# aGene_TCGA_df = aGene_TCGA_df
aGene_TCGA_df = aGene_TCGA_df.set_index('gene_symbol')
aGene_TCGA_df = aGene_TCGA_df.T
# aGene_TCGA_df = aGene_TCGA_df.reset_index()
print("Gene:", gene_symbol)
print(aGene_TCGA_df.head())
aGene_TCGA_cancer_id_df = pd.concat([aGene_TCGA_df, cancer_id_df],
join='inner',
axis=1)
# aGene_TCGA_cancer_id_df = pd.merge(aGene_TCGA_df,cancer_id_df,left_on='index', right_on='fullcode')
print(aGene_TCGA_cancer_id_df)
aGene_TCGA_cancer_id_df['Abbreviation'] = aGene_TCGA_cancer_id_df[
'Abbreviation'].astype('category')
cancer_types = aGene_TCGA_cancer_id_df['Abbreviation'].tolist()
cancer_types = list(set(cancer_types))
cancer_types = sorted(cancer_types)
## for other type boxplot
plt.figure(figsize=(10, 5))
# ax = sns.boxplot(x='Abbreviation', y=gene_symbol, data=aGene_TCGA_cancer_id_df, order=cancer_types, palette=pkmn_type_colors,
# showfliers=False);
ax = sns.boxplot(
x='Abbreviation',
y=gene_symbol,
data=aGene_TCGA_cancer_id_df,
order=cancer_types,
# palette=pkmn_type_colors,
color='w',
linewidth=1.5,
showfliers=False)
# ax = sns.swarmplot(x='Abbreviation', y=gene_symbol, data=aGene_TCGA_cancer_id_df, order=cancer_types, color=".25")
ax.set_title(gene_symbol)
ax.set_xlabel('TCGA')
ax.set_ylabel('Tumour vs Normal')
ax.xaxis.grid(True)
ax.set_xticklabels(ax.get_xticklabels(), rotation=45)
plt.setp(ax.artists, edgecolor='k', facecolor='w')
plt.setp(ax.lines, color='k')
file_dir = '/home/wch23/Project/LifeArc/TCGA/Result/boxplot/{}'.format(
cancer_type)
from pathlib import Path
Path(file_dir).mkdir(parents=True, exist_ok=True)
ax.get_figure().savefig(
'/home/wch23/Project/LifeArc/TCGA/Result/boxplot/{}/TCGA_boxplot_{}_original_new_v2.pdf'
.format(cancer_type, gene_symbol))
plt.show()
def get_sig_genes_by_anova(target_cancers, p_value_th):
tcga_cancer_diff_df = dh.load_obj(
"/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_diff_matrix_original"
)
with open(
'/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_types_has_normal.tsv'
) as cancer_type_f:
tcga_cancer_types = [x.strip() for x in cancer_type_f.readlines()]
cancer_id_addr = "/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/cancer_id_info.tsv"
cancer_id_df = pd.read_csv(cancer_id_addr, sep='\t')
# print(cancer_id_df)
# target_cancer_list=["PAAD"] # pancreas, ITGB5
target_cancer_list = target_cancers
# target_cancer_list = ["LUSC"] # Lung squamous cell carcinoma , Sox2
target_cancer = '_'.join(target_cancer_list)
cancer_id_df['Case_Ctrl'] = [
'Case' if x in target_cancer_list else 'Ctrl'
for x in cancer_id_df['Abbreviation']
]
# print(cancer_id_df)
#
tcga_cancer_diff_df = tcga_cancer_diff_df[
tcga_cancer_diff_df['gene_symbol'] != 'None']
tcga_cancer_diff_df = tcga_cancer_diff_df.drop(
columns=['ensembl_id', 'ensembl_gene'])
# # # print(tcga_cancer_diff_df.head())
# #
gene_value_mode = 'Diff'
# p_value_threshold = 1.0e-40 # -350 = significant gene(0) , -320 = significant(8300)
p_value_threshold = p_value_th
base_mean_threshold = 0
anova_result_addr = "/home/wch23/Project/LifeArc/TCGA/Result/TCGA_GTEx_original_{}_vs_other_{}_anova_result.csv".format(
target_cancer, gene_value_mode)
selected_genes_l = get_selected_genes_using_oneway_ANOVA(
tcga_cancer_diff_df, p_value_threshold, cancer_id_df, target_cancer,
gene_value_mode, anova_result_addr)
selected_genes_addr = "/home/wch23/Project/LifeArc/TCGA/Result/significant_genes_original_{}_from_{}_anova_{}.csv".format(
gene_value_mode, target_cancer, str(p_value_threshold))
sig_genes_df = pd.DataFrame(selected_genes_l, columns=["gene", 'pvalue'])
sig_genes_df.to_csv(selected_genes_addr, index=False)
tcga_id_mapping_addr = "/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_gtex_id_mapping.csv"
anova_result_with_ensembl_addr = "/home/wch23/Project/LifeArc/TCGA/Result/TCGA_GTEx_original_{}_vs_other_{}_anova_result_with_ensembl.csv".format(
target_cancer, gene_value_mode)
make_anova_result_with_symbol_ensembl(anova_result_addr,
tcga_id_mapping_addr,
anova_result_with_ensembl_addr)
################
# save significant genes
################
significant_genes_addr = "/home/wch23/Project/LifeArc/TCGA/Result/significant_genes_only_original_{}_from_{}_anova_{}.csv".format(
gene_value_mode, target_cancer, str(p_value_threshold))
anova_result_df = pd.read_csv(anova_result_addr)
sig_genes_df = anova_result_df[
anova_result_df['pvalue'] <= p_value_threshold]
# sig_genes = sig_genes_df['gene'].tolist()
#
# with open(significant_genes_addr, 'w') as sig_genes_f:
# sig_genes_f.write('\n'.join(sig_genes))
sig_genes_df.to_csv(significant_genes_addr, sep='\t')
def main_for_combine_genes():
tcga_cancer_diff_df = dh.load_obj(
"/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_diff_matrix_original"
)
with open(
'/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/tcga_cancer_types_has_normal.tsv'
) as cancer_type_f:
tcga_cancer_types = [x.strip() for x in cancer_type_f.readlines()]
cancer_id_addr = "/mnt/raid0_data/MTI_DATA/TCGA/GepiaTCGA/cancer_id_info.tsv"
cancer_id_df = pd.read_csv(cancer_id_addr, sep='\t')
# print(cancer_id_df)
# target_cancer_list=["PAAD"] # pancreas, ITGB5
# cancer_id_df['Case_Ctrl'] = ['Case' if x in target_cancer_list else 'Ctrl' for x in cancer_id_df['Abbreviation']]
# print(cancer_id_df)
#
tcga_cancer_diff_df = tcga_cancer_diff_df[
tcga_cancer_diff_df['gene_symbol'] != 'None']
tcga_cancer_diff_df = tcga_cancer_diff_df.drop(
columns=['ensembl_id', 'ensembl_gene'])
# print(cancer_id_df)
# gene_symbol_list = ["AC128709.3", "AL035258.1", "TCF4-AS1", "AC012498.2", "GSTA8P", "AC128709.2", "LINC01932",
# "POU6F2-AS2", "SCGB3A2", "KRT74", "AC022031.1", "LINC01206", "GBP6", "AC134043.2", "SERPINB13",
# "AC012498.1", "ADH7", "SFTPA1",
# "SFTPB"] # ['SRC', 'FYN', 'PRKCA', 'VTN', 'MYL12A','EPHA2' ,'SDC1','YES1']
# gene_symbol_list = ["CYTIP","B2M","AZI2","LINC00487","TNFSF12-TNFSF13","AL365203.2","HSPA1B"]
# gene_symbol_list = ["ITGA1","ITGA2","ITGA2B","ITGA3","ITGA4","ITGA5","ITGA6","ITGA7","ITGA8","ITGA9","ITGA10","ITGA11","ITGAD","ITGAE","ITGAL","ITGAM",
# "ITGAV","ITGAX","ITGB1","ITGB2","ITGB3","ITGB4","ITGB5","ITGB6","ITGB7","ITGB8","ITGBL1"]
# gene_symbol_list = ["TMEM52","EPB41L4B","AC011754.1","RBPJL","AC096633.1","PNLIP","CELP","LHFPL5","AC092535.1","TMED6"]
# gene_symbol_list = ["DLGAP1","DTNB","BHLHE40","PXN","CYLD","RYBP","GSC","LEFTY2","BMP7","NRP2","BMP4","OTX2","PRDM1","DENND2A","RAD51C","JARID2","KLF9","SOCS3","TGIF1","KANK1","MKRN1","FOS","DNMT3A","STAB2","RABIF","SNCG","ZIC3","FOXD3"]
gene_symbol_list = ["SOX2", "STAT1"]
gene_symbol = '_'.join(gene_symbol_list)
cancer_type = 'LUSC'
cancer_id_df = cancer_id_df.set_index('fullcode')
anova_result_df = pd.read_csv(
"/home/wch23/Project/LifeArc/TCGA/Result/TCGA_GTEx_original_{}_vs_other_Diff_anova_result.csv"
.format(cancer_type),
index_col=0)
aGene_TCGA_df = tcga_cancer_diff_df.loc[
tcga_cancer_diff_df['gene_symbol'].isin(gene_symbol_list)]
# aGene_TCGA_df = aGene_TCGA_df
aGene_TCGA_df = aGene_TCGA_df.set_index('gene_symbol')
aGene_TCGA_df = aGene_TCGA_df.T
# aGene_TCGA_df[gene_symbol] = aGene_TCGA_df.sum(axis=1)
aGene_TCGA_df[gene_symbol] = aGene_TCGA_df.mean(axis=1)
# aGene_TCGA_df = aGene_TCGA_df.reset_index()
aGene_TCGA_df = aGene_TCGA_df.drop(columns=gene_symbol_list)
print("Gene:", gene_symbol)
print(aGene_TCGA_df.head())
aGene_TCGA_cancer_id_df = pd.concat([aGene_TCGA_df, cancer_id_df],
join='inner',
axis=1)
# aGene_TCGA_cancer_id_df = pd.merge(aGene_TCGA_df,cancer_id_df,left_on='index', right_on='fullcode')
print(aGene_TCGA_cancer_id_df)
# anova_p_value = anova_result_df.loc[gene_symbol].values[0]
# print(anova_p_value)
aGene_TCGA_cancer_id_df['Abbreviation'] = aGene_TCGA_cancer_id_df[
'Abbreviation'].astype('category')
cancer_types = aGene_TCGA_cancer_id_df['Abbreviation'].tolist()
cancer_types = list(set(cancer_types))
cancer_types = sorted(cancer_types)
pkmn_type_colors = [
'#78C850', # Grass
'#F08030', # Fire
'#6890F0', # Water
'#A8B820', # Bug
'#A8A878', # Normal
'#A040A0', # Poison
'#F8D030', # Electric
'#E0C068', # Ground
'#EE99AC', # Fairy
'#C03028', # Fighting
'#F85888', # Psychic
'#B8A038', # Rock
'#705898', # Ghost
'#98D8D8', # Ice
'#7038F8', # Dragon
]
# ax = sns.boxplot(x='Abbreviation', y=gene_symbol, data=aGene_TCGA_cancer_id_df, order=cancer_types,
# showfliers=False);
# ax.set_title(gene_symbol)
# ax.set_xticklabels(ax.get_xticklabels(), rotation=90)
# ax.get_figure().savefig('/home/wch23/Project/LifeArc/TCGA/Result/boxplot/{}/TCGA_boxplot_{}_{}_original.pdf'.format(cancer_type,gene_symbol, anova_p_value))
## for other type boxplot
plt.figure(figsize=(10, 5))
# ax = sns.boxplot(x='Abbreviation', y=gene_symbol, data=aGene_TCGA_cancer_id_df, order=cancer_types, palette=pkmn_type_colors,
# showfliers=False);
ax = sns.boxplot(
x='Abbreviation',
y=gene_symbol,
data=aGene_TCGA_cancer_id_df,
order=cancer_types,
# palette=pkmn_type_colors,
color='w',
linewidth=1.5,
showfliers=False)
# ax = sns.swarmplot(x='Abbreviation', y=gene_symbol, data=aGene_TCGA_cancer_id_df, order=cancer_types, color=".25")
ax.set_title(gene_symbol)
ax.set_xlabel('TCGA')
ax.set_ylabel('Tumour vs Normal')
ax.xaxis.grid(True)
ax.set_xticklabels(ax.get_xticklabels(), rotation=45)
################
# this is for making color brighter
#################
# for patch in ax.artists:
# r, g, b, a = patch.get_facecolor()
# patch.set_facecolor((r, g, b, .3))
#########################
plt.setp(ax.artists, edgecolor='k', facecolor='w')
plt.setp(ax.lines, color='k')
ax.get_figure().savefig(
'/home/wch23/Project/LifeArc/TCGA/Result/boxplot/{}/TCGA_boxplot_{}_original_new_v2.pdf'
.format(cancer_type, gene_symbol
# anova_p_value
))
plt.show()