def tcell_read_metabolomics_frames():
proc_tcell_t = cache.TsvFileTracker(os.path.join(cache.get_cache_path(), metabolite_expression_name + ".tsv.gz"), tcell_read_metabolomics_data)
metabolomics_df = proc_tcell_t.read_file()
values,cols = [],[]
for coln in metabolomics_df.columns:
if "non act" in coln:
time = -1.
elif "ON" in coln:
time = 0.
elif "act 3h" in coln:
time = 3.
elif "act 12h" in coln:
time = 12.
elif "act 14h" in coln:
time = 14.
elif "act 24h" in coln:
time = 24.
elif "act 2d" in coln:
time = 48.
elif "act 3d" in coln:
time = 72.
elif "act 4d" in coln:
time = 96.
else:
print(coln)
dish = coln.split('-')[0]
rep = coln.split('-')[2].replace(" ","")
values += [time]
cols += ['_'.join([dish,str(int(time)),rep])]
phenotype_df = pd.DataFrame(columns=cols, data=[values], index=["Time"])
metabolomics_df.columns = cols
return phenotype_df, metabolomics_df
def generate_reactome_sunburst(values_df, sb_configuration = conf_human):
base = cache.get_cache_path()
json = generate_reactome_sunburst_json(values_df, sb_configuration)
write_json(json, os.path.join(base, 'results.json'))
resource_path = os.path.dirname(inspect.getsourcefile(run_server))
shutil.copy(os.path.join(resource_path,"index.html"), base)
shutil.copy(os.path.join(resource_path,"breadcrumb.js"), base)
run_server.run_sunburst(path=base)
def tcell_read_proteomics_data():
"""This function is quite convoluted as it downloads an excelfile from a publication and extracts a dataframe. The function also caches intermediate files"""
tcell_prot_xls = cache.UrlFileCache(os.path.join(cache.get_cache_path(), protein_expression_name + ".xlsx"),proteomics_data_url)
proteomics_df = pd.read_excel(tcell_prot_xls.get_file_name(), sheet_name = "Data", index_col=0, usecols="A,D:U")
# proteomics_df = pd.read_excel(tcell_prot_xls.get_file_name(), sheet_name = "Data", index_col=0, usecols="A,V:AM")
proteomics_df = proteomics_df - proteomics_df.mean() # Normalize by subtracting column mean
proteomics_df = proteomics_df.apply(np.exp2) # The excel data is in log2 space, return it to normal
proteomics_df = proteomics_df.groupby("Protein IDs", group_keys=False).apply(one_row_per_proteoform).reset_index(drop=True)
proteomics_df.set_index("ProteinID", inplace=True)
return proteomics_df
def get_reactome_df(organism = "HSA", gene_anot = "Ensembl"):
fn = gene_anot + reactome_fn
path = os.path.join(cache.get_cache_path(),fn)
url = reactome_url + fn
reactome_df = pd.read_csv(cache.download_file(path, url),
sep='\t',
header=None,
usecols=[0,1,3],
names=["gene","reactome_id","reactome_name"])
organism = "R-" + organism
reactome_df = reactome_df[reactome_df["reactome_id"].str.startswith(organism) ]
return reactome_df
def tcell_read_metabolomics_data():
"""This function is quite convoluted as it downloads an excelfile from a publication and extracts a dataframe, idexed by chebi. The function also caches intermediate files"""
tcell_metabol_xls = cache.UrlFileCache(os.path.join(cache.get_cache_path(), metabolite_expression_name + ".xlsx"), metabolomics_data_url)
metabolomics_df = pd.read_excel(tcell_metabol_xls.get_file_name(), sheet_name = "normalized by sample mean", index_col=0, usecols="A,C:HN", skiprows = [0])
#metabolomics_df = pd.read_excel(tcell_metabol_xls.get_file_name(), sheet_name = "normalized by sample mean", index_col=0, usecols="A,C:HN", skiprows = [0])
for col in metabolomics_df.columns:
# Average all technical replicates (Named by trailing ".1")
if len(col.split('.'))>1 and col.split('.')[1] == "1":
remcol = col.split('.')[0]
metabolomics_df[remcol] = scipy.stats.gmean(metabolomics_df[[remcol,col]],axis=1)
metabolomics_df.drop(col, axis=1, inplace=True)
metabolomics_df.index.name = "KEGG_ID"
metabolomics_df = metabolomics_df.apply(np.exp2) # The excel data is in log2 space, return it to normal
k = KEGG(verbose=False)
map_kegg_chebi = k.conv("chebi", "compound")
metabolomics_df = metabolomics_df.groupby("KEGG_ID", group_keys=False).apply(lambda x: one_row_per_compound_convert(x, map_kegg_chebi)).reset_index(drop=True)
metabolomics_df.set_index("MetaboliteID", inplace=True)
return metabolomics_df
def tcell_read_proteomics_frames():
proc_tcell_t = cache.TsvFileTracker(os.path.join(cache.get_cache_path(), protein_expression_name + ".tsv.gz"),tcell_read_proteomics_data)
proteomics_df = proc_tcell_t.read_file()
values,cols = [],[]
for coln in proteomics_df.columns:
if "notact" in coln:
time = 0.
elif "act12h" in coln:
time = 12.
elif "act24h" in coln:
time = 24.
elif "act48h" in coln:
time = 48.
elif "act72h" in coln:
time = 72.
elif "act96h" in coln:
time = 96.
else:
print(coln)
not_sure = coln.split('_')[0].replace("q","")
rep = int(coln.split('_')[3].replace(" ",""))
if rep<18:
dish = int(rep/5)+2
rep = rep%5+1
elif rep<26:
dish = rep%3+2
rep = 3
elif rep<31:
dish = (rep-1)%3+2
rep = 3
else:
dish = (rep-2)%3+2
rep = 3
values += [time]
cols += ['_'.join([str(dish),str(int(time)),str(rep)])]
proteomics_df.columns = cols
phenotype_df = pd.DataFrame(columns=cols, data=[values], index=["Time"])
return phenotype_df, proteomics_df
def generate_reactome_tree(sb_configuration = get_conf_human()):
""" Download the reactome pathway relation file and generate a tree structure """
# Derive a filename from the url
fn = os.path.basename(urlparse(url_to_reactome_relation_file).path)
path = os.path.join(cache.get_cache_path(),fn)
return generate_root_node(cache.download_file(path, url_to_reactome_relation_file), sb_configuration)