def get_sum_signs(rna,
mcff_args='-t 1',
mcff_timeout=300,
win_size=79,
win_skip=1):
'''Takes as an argument an RNA as a SeqRecord object.
Returns a vector of the sum of signatures for all windows (non-averaged)
and the number of windows used.
mcff_args = mcff parameters, defaulted to -t 1
mcff_timeout = time to take before mcff exits
win_size = length of windows to use
win_skip = how many nucleotides to move to the right each time'''
win_count = 0
signature_vector = np.zeros(940)
for i in range(0, len(rna.seq), win_skip):
# exits loop when the passed window is shorter than win_size
if len(rna.seq[i:i + win_size]) < win_size:
break
# calculates the signature for passed window
cur_win = rna.seq[i:i + win_size]
dotbs, shapes = mv.dotbs_and_shapes(cur_win,
parameters=mcff_args,
timeout_in_seconds=mcff_timeout)
signature_vector += mv.shape60_ncm40_ncmexp500_expseq340(
cur_win, dotbs, shapes)
win_count += 1
return (signature_vector, win_count)
def get_sum_signs_int(rna, mcff_args='-t 1', mcff_timeout=300, win_size=79, win_skip=WIN_SKIP):
'''Takes as an argument an RNA as a SeqRecord object.
Returns a list of vectors for the motif counts, along with the windows used and the number of dot-brackets used.
Also returns the rna_id of the processed transcript.
Used to make more lightweight .csv files in which all numeric entries are ints
mcff_args = mcff parameters, defaulted to -t 1
mcff_timeout = time to take before mcff exits
win_size = length of windows to use
win_skip = how many nucleotides to move to the right each time'''
win_count = 0
all_sign_vectors = []
dotbs_in_win = []
nt_position = []
for i in range(0, len(rna.seq), win_skip):
# exits loop when the passed window is shorter than win_size
if len(rna.seq[i:i+win_size]) < win_size:
break
# calculate the signature for passed window
cur_win = rna.seq[i:i+win_size]
dotbs, shapes = mv.dotbs_and_shapes(cur_win, parameters=mcff_args, timeout_in_seconds=mcff_timeout)
cur_signature_vector, nb_dotbs, _, _ = mv.shape60_ncm40_ncmexp500_expseq340_nodiv(cur_win, dotbs, shapes)
win_count += 1
all_sign_vectors.append(cur_signature_vector)
# keeps track of the dotbs and nt position
dotbs_in_win.append(nb_dotbs)
nt_position.append(i)
int_vectors = np.array( [arr.astype(int) for arr in all_sign_vectors] )
return (int_vectors, win_count, dotbs_in_win, rna.id, nt_position)
# retrieve the DF of motifs from file
print('Loading file...')
df = pd.read_parquet(
f'/u/floresj/mRNA_norm/mRNA_vectors/mrna_folded_int_subset{FILE_NUMBER}'
)
print('File has loaded.')
df_indices = sorted(list({rna_id for rna_id, _ in df.index.values}))
# calculate the frequency for aptamer-21
seq_list = [
seq for seq in parse(
'/u/floresj/Transcriptome_scanning_apta21/aptamer_21.fa',
format='fasta')
]
apt_21 = seq_list[0]
dotbs, shapes = mv.dotbs_and_shapes(apt_21.seq)
apt21_sign, _, _ = mv.shape60_ncm40_ncmexp500_expseq340(
apt_21.seq, dotbs, shapes)
## keep track of execution time
print('-------------------------------')
print('Version: Thursday July 26, 2018')
print('Time\t\tProcessed')
start_time = dt.now()
# multiprocessing setup
multiprocessing.set_start_method('spawn')
compteur = TimeCounter(len(df_indices))
pool = multiprocessing.Pool(20)
# process all transcripts