def visualize(prim_str, pred_string, true_string, save_imgs = False, suffix_img = ''):
corr_string = balance_op_tmp(pred_string)
print(f'pred: {pred_string}')
print(f"true: {true_string}")
print(f'corr: {corr_string}')
with open('tmp_vis_pred.txt', 'w') as f:
f.write(prim_str+ os.linesep)
f.write(corr_string+ os.linesep)
f.close
with open('tmp_vis_true.txt', 'w') as f:
f.write(prim_str+ os.linesep)
f.write(true_string+os.linesep)
f.close
plt.figure(figsize = (20, 20))
plt.title('predicted')
cg = forgi.load_rna('tmp_vis_pred.txt', allow_many=False)
fvm.plot_rna(cg, text_kwargs={"fontweight":"black"}, lighten=0.7,backbone_kwargs={"linewidth":3})
if save_imgs:
plt.savefig(f'pred_{suffix_img}.jpg')
plt.show()
plt.figure(figsize = (20, 20))
plt.title('original')
cg = forgi.load_rna('tmp_vis_true.txt', allow_many=False)
fvm.plot_rna(cg, text_kwargs={"fontweight":"black"}, lighten=0.7,backbone_kwargs={"linewidth":3})
plt.savefig(f'true_{suffix_img}.jpg')
plt.show()
def draw_2nd_structure(srna):
for i in range(3):
name = srna.loc[i, "name"]
sequence = srna.loc[i, "sequence"]
folding = fold_rna(srna.loc[i, "sequence"])[0]
print("##############", name, "###############")
print(sequence)
print(folding)
print()
fx_test = ">{0}\n{1}\n{2}\n".format(name, sequence, folding)
textfile = open('./resources/' + srna.loc[i, "name"] + '.fx', "w")
textfile.write(fx_test)
textfile.close()
for i in range(3):
##Print Structure
plt.figure(figsize=(20, 20))
cg = forgi.load_rna('./resources/' + srna.loc[i, "name"] + '.fx',
allow_many=False)
fvm.plot_rna(cg,
text_kwargs={"fontweight": "black"},
lighten=0.7,
backbone_kwargs={"linewidth": 3})
plt.show()
# plt.savefig(srna.loc[i, "name"]+'.png')
return srna
def create_pairing_matrix(seq, num_backtrack=10):
# create fold compound object
fc = RNA.fold_compound(seq, md)
# compute MFE
(ss, mfe) = fc.mfe()
# rescale Boltzmann factors according to MFE
fc.exp_params_rescale(mfe)
# compute partition function to fill DP matrices
fc.pf()
structures = []
pair_tables = []
pair_matrices = []
for s in fc.pbacktrack(num_backtrack):
structures.append(s)
bg, = forgi.load_rna(s)
pair_tables.append(bg.to_pair_table())
for pair_idx in pair_tables:
pair_matrix = np.eye(len(seq))
pair_idx = np.array(pair_idx[1:]) - 1
for i, j in enumerate(pair_idx):
if j == -1: # unpaired
continue
else:
pair_matrix[i, j] += 1
pair_matrix = pair_matrix / np.sum(pair_matrix, axis=-1, keepdims=True)
pair_matrices.append(pair_matrix)
pair_matrix = np.mean(pair_matrices, axis=0)
return pair_matrix
def load(self, pattern):
num_loaded = 0
num_selected = 0
if not isinstance(pattern, list):
pattern = [pattern]
fns = []
for pat in pattern:
fns.extend(glob(pat))
if len(fns) > 1000:
lev = logging.WARNING
else:
lev = logging.INFO
log.log(lev, "Loading %s files", len(fns))
with _LoggingContext(logging.getLogger(), logging.CRITICAL):
for fn in fns:
num_selected += 1
if fn not in self.cgs:
cgs = forgi.load_rna(fn)
if len(cgs) != 1:
raise ValueError(
"Expected 1 RNA component in file {}, found {}:{}".
format(fn, len(cgs)), [cg.name for cg in cgs])
cg, = cgs
cg.infos["filename"] = fn
self.cgs[fn] = cg
num_loaded += 1
if fn not in self.pdds:
points = []
try:
pd_pdd = pd.read_csv(fn + ".pdd.csv")
except:
for i in range(1, cg.seq_length + 1):
points.append(
cg.get_virtual_residue(
i, allow_single_stranded=True))
x, y = ftuv.pair_distance_distribution(
points, stepsize=self.stepsize)
df = pd.DataFrame({"step": x, "count": y})
df.to_csv(fn + ".pdd.csv")
else:
x = pd_pdd["step"]
y = pd_pdd["count"]
self.pdds[fn] = (x, y)
scores = self.get_scores(pattern)
if scores:
minsc = scores[0]
maxsc = scores[-1]
else:
minsc = None
maxsc = None
return num_selected, num_loaded, minsc, maxsc
def structure(sequence, dot_bracket, mirna_name, index):
out_dir = index + "/mirna"
with open(out_dir + "/" + mirna_name + ".db", "w") as temp_file:
temp_file.write(sequence + '\n')
temp_file.write(dot_bracket)
cg = forgi.load_rna(out_dir + "/" + mirna_name + ".db", allow_many=False)
fvm.plot_rna(cg,
text_kwargs={"fontweight": "black"},
lighten=0.7,
backbone_kwargs={"linewidth": 3})
plt.savefig("gui/src/assets/" + mirna_name + ".png")
def create_multiclass_vector(seq, num_backtrack=10):
# create fold compound object
fc = RNA.fold_compound(seq, md)
# compute MFE
(ss, mfe) = fc.mfe()
# rescale Boltzmann factors according to MFE
fc.exp_params_rescale(mfe)
# compute partition function to fill DP matrices
fc.pf()
structures = []
dict_map = {'s': 'P', 'i': 'I', 'h': 'H', 'm': 'M', 'f': 'E', 't': 'E'}
one_hot_vector = np.zeros((len(seq), 5))
for s in fc.pbacktrack(num_backtrack):
structures.append(s)
bg, = forgi.load_rna(s)
element_string = bg.to_element_string()
element_string = ''.join(map(lambda x: dict_map[x], element_string))
one_hot_vector += one_hot_encode_struct(element_string,
method='signed')
one_hot_vector /= 1.0 * num_backtrack
return one_hot_vector
def plot_rna_struct(seq,
struct,
ax=None,
offset=(0, 0),
text_kwargs={},
backbone_kwargs={},
basepair_kwargs={},
highlight_bp_idx=[],
highlight_nt_idx=[],
lighten=0.7,
saveto='tmp.png'):
with open('tmp.fa', 'w') as file:
file.write('>tmp\n%s\n%s' % (seq, struct))
cg = forgi.load_rna('tmp.fa', allow_many=False)
RNA.cvar.rna_plot_type = 1
fig = plt.figure(figsize=(30, 30))
coords = []
bp_string = cg.to_dotbracket_string()
if ax is None:
ax = plt.gca()
if offset is None:
offset = (0, 0)
elif offset is True:
offset = (ax.get_xlim()[1], ax.get_ylim()[1])
else:
pass
vrna_coords = RNA.get_xy_coordinates(bp_string)
# TODO Add option to rotate the plot
for i, _ in enumerate(bp_string):
coord = (offset[0] + vrna_coords.get(i).X,
offset[1] + vrna_coords.get(i).Y)
coords.append(coord)
coords = np.array(coords)
# First plot backbone
bkwargs = {"color": "grey", "zorder": 0, "linewidth": 0.5}
bkwargs.update(backbone_kwargs)
ax.plot(coords[:, 0], coords[:, 1], **bkwargs)
# Now plot basepairs
basepairs_hl, basepairs_nonhl = [], []
for s in cg.stem_iterator():
for p1, p2 in cg.stem_bp_iterator(s):
if (p1 - 1, p2 - 1) in highlight_bp_idx:
basepairs_hl.append([coords[p1 - 1], coords[p2 - 1]])
else:
basepairs_nonhl.append([coords[p1 - 1], coords[p2 - 1]])
if len(basepairs_hl) > 0:
basepairs_hl = np.array(basepairs_hl)
bpkwargs_hl = {"color": 'red', "zorder": 0, "linewidth": 3}
bpkwargs_hl.update(basepair_kwargs)
ax.plot(basepairs_hl[:, :, 0].T, basepairs_hl[:, :, 1].T,
**bpkwargs_hl)
if len(basepairs_nonhl) > 0:
basepairs_nonhl = np.array(basepairs_nonhl)
bpkwargs_nonhl = {"color": 'black', "zorder": 0, "linewidth": 0.5}
bpkwargs_nonhl.update(basepair_kwargs)
ax.plot(basepairs_nonhl[:, :, 0].T, basepairs_nonhl[:, :, 1].T,
**bpkwargs_nonhl)
# Now plot circles
for i, coord in enumerate(coords):
if i in highlight_nt_idx:
c = 'green'
h, l, s = colorsys.rgb_to_hls(*mc.to_rgb(c))
if lighten > 0:
l += (1 - l) * min(1, lighten)
else:
l += l * max(-1, lighten)
c = colorsys.hls_to_rgb(h, l, s)
circle = plt.Circle((coord[0], coord[1]),
edgecolor="black",
facecolor=c)
else:
circle = plt.Circle((coord[0], coord[1]),
edgecolor="black",
facecolor="white")
ax.add_artist(circle)
if cg.seq:
if "fontweight" not in text_kwargs:
text_kwargs["fontweight"] = "bold"
ax.annotate(cg.seq[i + 1],
xy=coord,
ha="center",
va="center",
**text_kwargs)
all_coords = list(coords)
ntnum_kwargs = {"color": "gray"}
ntnum_kwargs.update(text_kwargs)
for nt in range(10, cg.seq_length, 10):
# We try different angles
annot_pos = _find_annot_pos_on_circle(nt, all_coords, cg)
if annot_pos is not None:
ax.annotate(str(nt),
xy=coords[nt - 1],
xytext=annot_pos,
arrowprops={
"width": 1,
"headwidth": 1,
"color": "gray"
},
ha="center",
va="center",
zorder=0,
**ntnum_kwargs)
all_coords.append(annot_pos)
datalim = ((min(list(coords[:, 0]) + [ax.get_xlim()[0]]),
min(list(coords[:, 1]) + [ax.get_ylim()[0]])),
(max(list(coords[:, 0]) + [ax.get_xlim()[1]]),
max(list(coords[:, 1]) + [ax.get_ylim()[1]])))
ax.set_aspect('equal', 'datalim')
ax.update_datalim(datalim)
ax.autoscale_view()
ax.set_axis_off()
plt.savefig(saveto, dpi=350)
plt.close(fig)
import RNA
import forgi
import numpy as np
import pandas as pd
import pickle
count = 0
to_pickle = []
pos_dotbracket = open('pos_dotbracket_FXR1.txt', 'r')
for dotbracket in pos_dotbracket:
source = []
target = []
bg, = forgi.load_rna(dotbracket.rstrip())
pt = bg.to_pair_table()[1:]
#am = np.zeros((100,100))
for i in range(len(pt)):
if i + 1 < len(pt):
source.append(i)
target.append(i + 1)
if i - 1 > -1:
source.append(i)
target.append(i - 1)
if pt[i] != 0:
source.append(i)
target.append(pt[i] - 1)
source = np.array(source)
target = np.array(target)
to_pickle.append(np.stack((source, target)))
if count % 1000 == 0:
print(count)
count += 1
# print(row[0])
seq = row[2]
# print('seq')
# print(seq)
dotbracket = row[14]
# print("dotbracket:")
# print(dotbracket)
if len(seq) != len(dotbracket):
print('Different Lengths, failed:')
print(row)
continue
tmp_file = open("tmp.txt", "w")
tmp_file.write(seq + '\n')
tmp_file.write(dotbracket + '\n')
tmp_file.close()
cg = forgi.load_rna('tmp.txt', allow_many=False)
result = fgb.BulgeGraph.to_element_string(cg, with_numbers=True)
result = result.splitlines()
row.append(str(result[0]))
row.append(str(result[1]))
writeCSV.writerow(row)
# all_col.append(row)
# writeCSV.writerows(all_col)
# bg = fgb.BulgeGraph.from_dotbracket('((..))..((..))')
# rna_seq_dotb = ['CGCUUCAUAUAAUCCUAAUGAUAUGGUUUGGGAGUUUCUACCAAGAGCCUUAAACUCUUGAUUAUGAAGUU', '((((((((((..(((((((.......)))))))......).(((.((.......))))))..)))))))).']
# tmp_file = open("tmp.txt","w")
# tmp_file.write(rna_seq_dotb[0]+ '\n')
# tmp_file.write(rna_seq_dotb[1]+ '\n')
# tmp_file.writelines(rna_seq_dotb)
def search_dot_pair_stem(args):
dot_dict = util.read_dot(dot=args.dot)
import forgi
d = dot_dict['dotbracket']
s = '...(((((...[[[.))))).((((((((((.(((((((((.....(((.(((..((...(((....((..........))...)))))......(((......((((..((..((....(((..................((((....(((((((.....))))))).....)))).......((((...((((((....))))))...))))....((((((.......(((((.((((...((((.((((((((....))))))))..)))).)))).....)))))......))))))...........((((.((((......))))))))....)))...))))..))))(((..(.(((....((((((((.......))))))))))).....))))...((((((((....))))...))))))).((((((..........)))))).((((....))))...)))))).).....(.(((...(((((...))))).)))).)).))))))....(((((((((((((....))).))))))).)))......(((.(((.......)))).)).........(((((((((....[[[[.....[[.)))]].......]]]])))))).))))))))))..........(((((.....((((...(((.......(((.(((((((((((((.((((....))))....))))))))..)))))))).......((((.(((((...(((((((......)))))))....)))))))))................................................................................................................................(((((((((..(((((((((..((((((((...(((......)))......))))))))..))....(..((....)))))))))).))))).))))...)))...))))....((((((...((...((((.........))))...))))))))..........[[[[[[.(((..((((((((.(((((....)))))))))))))..)))...[[..))]]...]]]....]]].)))..(((.....((((....))))....)))...]]]..(((((.(((((((..((..(((((((((((((((((....((((........))))........(((((((....(((((........((((((........))))))......)))))...((.((((..(((((((((...(((((((((....)))..((((......))))..)))))).....((((.(((.((((..((((....(((..((((....)))).)))....))))..)))))))..((((((((.....))))))))....))))...)))).)))...).))))))).....)))))))...)).))))))))))...(((((((.....(((.......((..((((....))))..)).....))).....)))))))......(...((((((((........))))))))...).....))))).....((((((((.......))))))))......))...)))))))))).))....((.((.(.((((((((.((.((((((((((((..(((((((((((((((.((((((((((((.....))))))))))))...)))))))))))))))..))))))))))))).)))))))))..).))..))....((((((((((....))))))))))........'
bg, = forgi.load_rna(s)
pair_dict = nested_dict(2, list)
reg_str = "[\(\)]{" + str(args.min_len) + ",}"
# for n,match in enumerate(re.finditer(reg_str, dot_dict['dotbracket'])):
for n, match in enumerate(re.finditer(reg_str, s)):
# if n >= 3: continue
start = match.span()[0] # 0-based
end = match.span()[1] # # 0-based, not include
start_pair = bg.pairing_partner(start + 1) # 1-based
end_pair = bg.pairing_partner(end) # 1-based
pos_ls = [start, end, start_pair, end_pair]
# print(pos_ls)
if None in pos_ls: continue
middle = int((max(pos_ls) + min(pos_ls)) / 2)
mask_len = max(pos_ls) - min(pos_ls)
if mask_len > 100: continue
fragment_start = middle - 50
fragment_end = middle + 50
if fragment_start < 0: continue
if fragment_end > len(dot_dict['dotbracket']): continue
print(n, match.start(), match.span(), match.group(), start_pair,
end_pair, middle, fragment_start, fragment_end)
if (end - start) != (abs(start_pair - end_pair) + 1): continue
pair_start = min(start_pair, end_pair) - 1
pair_end = max(start_pair, end_pair)
if end < pair_start:
gap_start = end
gap_end = pair_start
else:
gap_start = pair_end
gap_end = start
gap_dot = s[pair_end:start]
if len(set(gap_dot)) != 1: continue
pair_dict[n]['gap_start'] = gap_start
pair_dict[n]['gap_end'] = gap_end
pair_dict[n]['stem_start'] = start
pair_dict[n]['stem_end'] = end
pair_dict[n]['pair_start'] = pair_start
pair_dict[n]['pair_end'] = pair_end
pair_dict[n]['fragment_start'] = fragment_start
pair_dict[n]['fragment_end'] = fragment_end
generate_mask_region_validate(
shape_out=args.shape_out,
tx=args.tx,
species=args.species,
mask_start=start,
mask_end=end,
fragment_start=fragment_start,
fragment_end=fragment_end,
savefn_dir=
'/home/gongjing/project/shape_imputation/data/hek_wc_vivo_rRNA/3.shape/mask_specific_regions/ss_pair',
plot_gradient=1)
generate_mask_region_validate(
shape_out=args.shape_out,
tx=args.tx,
species=args.species,
mask_start=pair_start,
mask_end=pair_end,
fragment_start=fragment_start,
fragment_end=fragment_end,
savefn_dir=
'/home/gongjing/project/shape_imputation/data/hek_wc_vivo_rRNA/3.shape/mask_specific_regions/ss_pair',
plot_gradient=1)
pair_df = pd.DataFrame.from_dict(pair_dict, orient='index')
savefn = '/home/gongjing/project/shape_imputation/data/hek_wc_vivo_rRNA/3.shape/mask_specific_regions/ss_pair/stem_pair.txt'
pair_df.to_csv(savefn, header=True, index=True, sep='\t')