def run(args):
# Set logging level.
logging_debug_opt = False
LOGGER.addHandler(create_logging_handler(logging_debug_opt))
LOGGER.setLevel(logging.DEBUG)
LOGGER.info("Using configuration {}.".format(args.config_filename))
cfg = ConfigParser()
cfg.read(args.config_filename)
in_fname = cfg['data']['modules'] if not args.input else args.input
LOGGER.info("Loading modules from {}.".format(in_fname))
# Loading from YAML is extremely slow. Therefore this is a potential performance improvement.
# Potential improvements are switching to JSON or to use a CLoader:
# https://stackoverflow.com/questions/27743711/can-i-speedup-yaml
if in_fname.endswith('.yaml'):
modules = load_from_yaml(in_fname)
else:
with open(in_fname, 'rb') as f:
modules = pickle.load(f)
# Filter out modules with to few genes.
min_genes = int(cfg['parameters']['min_genes'])
modules = list(filter(lambda m: len(m) >= min_genes, modules))
LOGGER.info("Loading databases.")
def name(fname):
return os.path.splitext(os.path.basename(fname))[0]
db_fnames = list(mapcat(glob.glob, cfg['data']['databases'].split(";")))
dbs = [
RankingDatabase(fname=fname, name=name(fname)) for fname in db_fnames
]
LOGGER.info("Calculating regulons.")
motif_annotations_fname = cfg['data']['motif_annotations']
mode = cfg['parameters']['mode']
with ProgressBar() if mode == "dask_multiprocessing" else NoProgressBar():
df = prune2df(dbs,
modules,
motif_annotations_fname,
rank_threshold=int(cfg['parameters']['rank_threshold']),
auc_threshold=float(cfg['parameters']['auc_threshold']),
nes_threshold=float(cfg['parameters']['nes_threshold']),
client_or_address=mode,
module_chunksize=cfg['parameters']['chunk_size'],
num_workers=args.num_workers)
LOGGER.info("Writing results to file.")
df.to_csv(cfg['parameters']['output'] if not args.output else args.output)
cor_p_thr = 0.001
memory_limit = 50e9
n_cells = 600
grouping_variable = 'celltype'
if __name__ == '__main__':
if sum(metadata.columns == grouping_variable) < 1:
exit('Grouping variable not found in metadata.')
## Load randing databases
db_fnames = glob.glob(db_folder)
def name(fname):
return os.path.basename(fname).split(".")[0]
dbs = [RankingDatabase(fname=fname, name=name(fname)) for fname in db_fnames]
dbs
## Initialize cluster
local_cluster = LocalCluster(n_workers=n_cores, threads_per_worker=1, processes=False, memory_limit=memory_limit)
custom_client = Client(local_cluster)
## Load TFs
tf_names = load_tf_names(TFs_file)
## Collect here regulons passing correlation filter
cortest_passed_regulons = []
def db():
return RankingDatabase(TEST_DATABASE_FNAME, TEST_DATABASE_NAME)
def db():
return RankingDatabase(TEST_DATABASE_FNAME, TEST_DATABASE_NAME,
NOMENCLATURE)
#---------------functions-----------------
def name(fname):
return os.path.splitext(os.path.basename(fname))[0]
#################################################################
##1. load df (already loaded in setup)
#df_cnt=pd.read_csv(f_in, index_col=0)
#2. tf genes
tf_name=load_tf_names(f_tf)
#3. ranking databases (only 2 mm10 dbs)
l_fname=list(Path(fd_db).glob('*.feather'))
l_db=[RankingDatabase(fname=fname, name=name(fname)) for fname in l_fname]
#3. run
if __name__ =='__main__':
# #1. Inference of co-expression modules
# print('Inference...')
# df_adj=grnboost2(df_cnt, tf_names=tf_name, verbose=True)
# df_adj.to_csv(f'{fd_out}/adj.csv', index=False)
#2. prune
df_adj=pd.read_csv(f'{fd_out}/adj.csv') #if missing, always stuck at 98%
print('Prune...')
l_mod=list(modules_from_adjacencies(df_adj, df_cnt))
with ProgressBar():
df_prune = prune2df(l_db, l_mod, f_motif)
auc_mtx = aucell(expr, regulons, num_workers=ppn)
tfs = [tf.strip('(+)') for tf in auc_mtx.columns]
auc_mtx.to_csv('{}/{}_auc_mtx.csv'.format(out_path, prefix))
print('finished calculation for %s' % (prefix))
if __name__ == "__main__":
wkdir = '/bgfs/alee/chelsea/projects/10X/AL1/codes'
os.chdir(wkdir)
db = [
RankingDatabase(
fname=
'../data/pySCENIC/ref/hg38__refseq-r80__10kb_up_and_down_tss.mc9nr.feather',
name='hg38__refseq-r80__10kb_up_and_down_tss.mc9nr')
]
tf_names = load_tf_names('../data/pySCENIC/ref/hs_hgnc_curated_tfs.txt')
CellTypes = [
'MCF7', 'T47D WT', 'T47D KO', 'MM134', 'SUM44', 'BCK4', 'MCF10A',
'HEK293'
]
for cell in CellTypes:
tmp = adata_raw[adata_raw.obs['CellType'] == cell]
RawSplicedCts = pd.DataFrame(tmp.layers['spliced'].todense(),
index=tmp.obs.index,
columns=tmp.var.index) # cell X gene
