pop_mgi = set(dfQ['MGI ID'].dropna().tolist())
pop_uniprot = set(dfQ['UniProtKB/Swiss-Prot ID'].dropna().tolist())
pop = pop_mgi.union(pop_uniprot)
elif layer == 'HS':
pop_uniprot = set(dfQ['UniProtKB/Swiss-Prot ID'].dropna().tolist())
pop = pop_uniprot
# Load GO
print("Load GO files")
godag = obo_parser.GODag(ontology)
# GAF files for both MM and DM
gaf_species = GafReader(name='GAF ' + layer + ' Specie', filename=annotation, godag=godag, namespaces=set(['BP']))
gaf_reactome = GafReader(name='GAF ' + layer + ' Reactome', filename=annotation_reactome, godag=godag, namespaces=set(['BP']))
# Dict of Associations
ns2assoc_species = gaf_species.get_ns2assc()
n_assoc_species = sum([len(v) for k, v in ns2assoc_species['BP'].items()])
print('Specie associations: {n:d}'.format(n=n_assoc_species))
# We also need to add the multi-species annotations
ns2assoc_reactome = gaf_reactome.get_ns2assc()
n_assoc_reactome = sum([len(v) for k, v in ns2assoc_reactome['BP'].items()])
print('Reactome associations: {n:d}'.format(n=n_assoc_reactome))
# combine associations
ns2assoc_combined = merge(ns2assoc_species, ns2assoc_reactome)
n_assoc_combined = sum([len(v) for k, v in ns2assoc_combined['BP'].items()])
print('Combined associations: {n:d}'.format(n=n_assoc_combined))
"""
# Use PANTHER HMM for HS?
layer = 'DM'
threshold_str = str(threshold).replace('.', 'p')
# GO Information
dict_annotation_file = {
'HS': 'goa_human.gaf',
'MM': 'mgi.gaf',
'DM': 'fb.gaf'
}
annotation = '../data/GeneOntology/' + dict_annotation_file[layer]
ontology = '../data/GeneOntology/go-basic.obo'
#
godag = obo_parser.GODag(ontology)
gaf = GafReader(name='GAF ' + layer, filename=annotation, godag=godag)
# Dict of Associations
ns2assoc = gaf.get_ns2assc()
# Load Population of Genes (for background comparison)
rFPKMFile = '../02-core_genes/results/FPKM/{layer:s}/{layer:s}-FPKM-{celltype:s}.csv.gz'.format(
celltype=celltype, layer=layer)
dfP = pd.read_csv(rFPKMFile, usecols=['id_gene', 'gene'])
# Load PCA
rPCAFile = 'results/pca/{celltype:s}/pca-{celltype:s}-{network:s}-{threshold:s}-{layer:s}-dim.csv.gz'.format(
celltype=celltype,
network=network,
threshold=threshold_str,
layer=layer)
df_pca = pd.read_csv(rPCAFile, index_col=0, encoding='utf-8')
# Population of genes (background) to test against