def test_anisotropy_planar(self):
# The anisotropy for planar symmetric objects converges to 1/4.
# See references 36-40 in doi:10.1063/1.4788616
planar = np.array([[0., 0, 1], [0, 1, 0], [0, -1, 0], [0, 0, 1],
[0, 0, 0], [0, 0.6, 0.6], [0, -0.6, 0.6],
[0, 0.6, -0.6], [0, -0.6, -0.6]])
self.assertAlmostEqual(ftmd.anisotropy(planar), 0.25, places=2)
def _get_descriptor(self, descriptor, domain=None):
"""
:param ensemble: an Ensemble or EnsembleView object
:param descriptor: A STRING. One of AVAILABLE_DESCRIPTORS
:param domain: An iterable of cg element names or None (whole cg)
:returns: A np.array
"""
if descriptor not in self.AVAILABLE_DESCRIPTORS:
raise ValueError("Descriptor {} not available.".format(descriptor))
if descriptor == "rog":
if domain:
return np.array([
ftmd.radius_of_gyration(
cg.get_poss_for_domain(domain, "vres")) for cg in self
])
else:
return np.array([cg.radius_of_gyration() for cg in self])
elif descriptor == "anisotropy":
if domain:
return np.array([
ftmd.anisotropy(cg.get_poss_for_domain(domain, "vres"))
for cg in self
])
else:
return np.array([
ftmd.anisotropy(cg.get_ordered_stem_poss()) for cg in self
])
elif descriptor == "asphericity":
if domain:
return np.array([
ftmd.asphericity(cg.get_poss_for_domain(domain, "vres"))
for cg in self
])
else:
return np.array([
ftmd.asphericity(cg.get_ordered_stem_poss()) for cg in self
])
def _get_descriptor(self, descriptor, domain=None):
"""
:param ensemble: an Ensemble or EnsembleView object
:param descriptor: A STRING. One of AVAILABLE_DESCRIPTORS
:param domain: An iterable of cg element names or None (whole cg)
:returns: A np.array
"""
if descriptor not in self.AVAILABLE_DESCRIPTORS:
raise ValueError("Descriptor {} not available.".format(descriptor))
if descriptor == "rog":
if domain:
return np.array([ftmd.radius_of_gyration(cg.get_poss_for_domain(domain, "vres")) for cg in self])
else:
return np.array([cg.radius_of_gyration() for cg in self])
elif descriptor == "anisotropy":
if domain:
return np.array([ftmd.anisotropy(cg.get_poss_for_domain(domain, "vres")) for cg in self])
else:
return np.array([ftmd.anisotropy(cg.get_ordered_stem_poss()) for cg in self])
elif descriptor == "asphericity":
if domain:
return np.array([ftmd.asphericity(cg.get_poss_for_domain(domain, "vres")) for cg in self])
else:
return np.array([ftmd.asphericity(cg.get_ordered_stem_poss()) for cg in self])
def anisotropy(cgs):
ai = np_nans(len(cgs))
for i, cg in enumerate(cgs):
ai[i] = ftmd.anisotropy(cg.get_ordered_stem_poss())
return ai
def describe_rna(cg, file_num, dist_pais, angle_pairs):
data = {}
data["nt_length"] = cg.seq_length
data["num_cg_elems"] = len(cg.defines)
for letter in "smifth":
data["num_" + letter] = len([x for x in cg.defines if x[0] == letter])
multiloops = cg.find_mlonly_multiloops()
descriptors = []
junct3 = 0
junct4 = 0
reg = 0
pk = 0
op = 0
for ml in multiloops:
descriptors = cg.describe_multiloop(ml)
if "regular_multiloop" in descriptors:
if len(ml) == 3:
junct3 += 1
elif len(ml) == 4:
junct4 += 1
reg += 1
if "pseudoknot" in descriptors:
pk += 1
if "open" in descriptors:
op += 1
data["3-way-junctions"] = junct3
data["4-way-junctions"] = junct4
#print (descriptors)
data["open_mls"] = op
# print(data["open_mls"][-1])
data["pseudoknots"] = pk
data["regular_mls"] = reg
data["total_mls"] = len(multiloops)
try:
data["longest_ml"] = max(len(x) for x in multiloops)
except ValueError:
data["longest_ml"] = 0
try:
data["rog_fast"] = cg.radius_of_gyration("fast")
except (ftmc.RnaMissing3dError, AttributeError):
data["rog_fast"] = float("nan")
data["rog_vres"] = float("nan")
data["anisotropy_fast"] = float("nan")
data["anisotropy_vres"] = float("nan")
data["asphericity_fast"] = float("nan")
data["asphericity_vres"] = float("nan")
else:
data["rog_vres"] = cg.radius_of_gyration("vres")
data["anisotropy_fast"] = ftmd.anisotropy(cg.get_ordered_stem_poss())
data["anisotropy_vres"] = ftmd.anisotropy(
cg.get_ordered_virtual_residue_poss())
data["asphericity_fast"] = ftmd.asphericity(cg.get_ordered_stem_poss())
data["asphericity_vres"] = ftmd.asphericity(
cg.get_ordered_virtual_residue_poss())
for from_nt, to_nt in dist_pairs:
try:
dist = ftuv.vec_distance(
cg.get_virtual_residue(int(from_nt), True),
cg.get_virtual_residue(int(to_nt), True))
except Exception as e:
dist = float("nan")
log.warning(
"%d%s File %s: Could not calculate distance between "
"%d and %d: %s occurred: %s", file_num, {
1: "st",
2: "nd",
3: "rd"
}.get(file_num % 10 * (file_num % 100 not in [11, 12, 13]),
"th"), cg.name, from_nt, to_nt,
type(e).__name__, e)
data["distance_{}_{}".format(from_nt, to_nt)] = dist
for elem1, elem2 in angle_pairs:
try:
angle = ftuv.vec_angle(cg.coords.get_direction(elem1),
cg.coords.get_direction(elem2))
except Exception as e:
angle = float("nan")
log.warning(
"%d%s File %s: Could not calculate angle between "
"%s and %s: %s occurred: %s", file_num, {
1: "st",
2: "nd",
3: "rd"
}.get(file_num % 10 * (file_num % 100 not in [11, 12, 13]),
"th"), cg.name, elem1, elem2,
type(e).__name__, e)
data["angle_{}_{}".format(elem1, elem2)] = angle
data["missing_residues_5prime"] = (len(cg.seq.with_missing[:1]) - 1)
data["missing_residues_3prime"] = (
len(cg.seq.with_missing[cg.seq_length:]) - 1)
data["missing_residues_middle"] = (
len(cg.seq.with_missing[1:cg.seq_length]) -
len(cg.seq[1:cg.seq_length]))
data["missing_residues_total"] = (len(cg.seq.with_missing[:]) -
len(cg.seq[:]))
fp = len(cg.seq.with_missing[:1]) - 1
tp = 0
old_bp = None
bp = None
for bp in cg.backbone_breaks_after:
fp += len(cg.seq.with_missing[bp:bp + 1].split('&')[1]) - 1
tp += len(cg.seq.with_missing[bp:bp + 1].split('&')[0]) - 1
tp += len(cg.seq.with_missing[cg.seq_length:]) - 1
data["missing_residues_5prime_chain"] = (fp)
data["missing_residues_3prime_chain"] = (tp)
data["missing_residues_middle_chain"] = (data["missing_residues_total"] -
fp - tp)
incomplete_elem_types = Counter(x[0] for x in cg.incomplete_elements)
data["s_with_missing"] = incomplete_elem_types["s"]
data["i_with_missing"] = incomplete_elem_types["i"]
data["m_with_missing"] = incomplete_elem_types["m"]
data["h_with_missing"] = incomplete_elem_types["h"]
mp = ""
if incomplete_elem_types["s"]:
for elem in cg.incomplete_elements:
if elem[0] != "s":
continue
for i in range(cg.defines[elem][0], cg.defines[elem][1]):
left_s = cg.seq.with_missing[i:i + 1]
if len(left_s) > 2:
right_s = cg.seq.with_missing[cg.pairing_partner(i + 1):cg.
pairing_partner(i)]
if len(right_s) > 2:
mp += "{}&{};".format(left_s, right_s)
data["missing_basepairs"] = mp
return data
def test_anisotropy_star(self):
star = np.array([[0., 0, 1], [0, 1, 0], [1, 0, 0], [-1, 0, 0],
[0, -1, 0], [0, 0, 1], [0, 0, 0]])
self.assertLessEqual(ftmd.anisotropy(star), 0.2)
def test_anisotropy_linear(self):
linear = np.array([[1., 1., 1.], [0., 0., 0.], [-1., -1., -1.],
[-2, -2, -2]])
self.assertAlmostEqual(ftmd.anisotropy(linear), 1)
def update(self, sm, step):
rog = ftur.anisotropy(sm.bg.get_ordered_stem_poss())
self.history[0].append(rog)
return "{:6.2f}".format(rog)
def test_anisotropy_no_coords(self):
a = np.array([])
self.assertTrue(np.isnan(ftmd.anisotropy(a)))
def describe_rna(cg, file_num, dist_pais, angle_pairs):
data = {}
data["nt_length"] = cg.seq_length
data["num_cg_elems"] = len(cg.defines)
for letter in "smifth":
data["num_" + letter] = len([x for x in cg.defines if x[0] == letter])
multiloops = cg.find_mlonly_multiloops()
descriptors = []
junct3 = 0
junct4 = 0
reg = 0
pk = 0
op = 0
for ml in multiloops:
descriptors = cg.describe_multiloop(ml)
if "regular_multiloop" in descriptors:
if len(ml) == 3:
junct3 += 1
elif len(ml) == 4:
junct4 += 1
reg += 1
if "pseudoknot" in descriptors:
pk += 1
if "open" in descriptors:
op += 1
data["3-way-junctions"] = junct3
data["4-way-junctions"] = junct4
#print (descriptors)
data["open_mls"] = op
#print(data["open_mls"][-1])
data["pseudoknots"] = pk
data["regular_mls"] = reg
data["total_mls"] = len(multiloops)
try:
data["longest_ml"] = max(len(x) for x in multiloops)
except ValueError:
data["longest_ml"] = 0
try:
data["rog_fast"] = cg.radius_of_gyration("fast")
except (ftmc.RnaMissing3dError, AttributeError):
data["rog_fast"] = float("nan")
data["rog_vres"] = float("nan")
data["anisotropy_fast"] = float("nan")
data["anisotropy_vres"] = float("nan")
data["asphericity_fast"] = float("nan")
data["asphericity_vres"] = float("nan")
else:
data["rog_vres"] = cg.radius_of_gyration("vres")
data["anisotropy_fast"] = ftmd.anisotropy(cg.get_ordered_stem_poss())
data["anisotropy_vres"] = ftmd.anisotropy(
cg.get_ordered_virtual_residue_poss())
data["asphericity_fast"] = ftmd.asphericity(cg.get_ordered_stem_poss())
data["asphericity_vres"] = ftmd.asphericity(
cg.get_ordered_virtual_residue_poss())
for from_nt, to_nt in dist_pairs:
try:
dist = ftuv.vec_distance(
cg.get_virtual_residue(int(from_nt), True),
cg.get_virtual_residue(int(to_nt), True))
except Exception as e:
dist = float("nan")
log.warning(
"%d%s File %s: Could not calculate distance between "
"%d and %d: %s occurred: %s", file_num, {
1: "st",
2: "nd",
3: "rd"
}.get(file_num % 10 * (file_num % 100 not in [11, 12, 13]),
"th"), cg.name, from_nt, to_nt,
type(e).__name__, e)
data["distance_{}_{}".format(from_nt, to_nt)] = dist
for elem1, elem2 in angle_pairs:
try:
angle = ftuv.vec_angle(cg.coords.get_direction(elem1),
cg.coords.get_direction(elem2))
except Exception as e:
angle = float("nan")
log.warning(
"%d%s File %s: Could not calculate angle between "
"%s and %s: %s occurred: %s", file_num, {
1: "st",
2: "nd",
3: "rd"
}.get(file_num % 10 * (file_num % 100 not in [11, 12, 13]),
"th"), cg.name, elem1, elem2,
type(e).__name__, e)
data["angle_{}_{}".format(elem1, elem2)] = angle
return data
def test_anisotropy_no_coords(self):
a = np.array([])
self.assertTrue(np.isnan(ftmd.anisotropy(a)))
def test_anisotropy_planar(self):
# The anisotropy for planar symmetric objects converges to 1/4.
# See references 36-40 in doi:10.1063/1.4788616
planar = np.array([[0., 0, 1], [0, 1, 0], [0, -1, 0], [0, 0, 1], [0, 0, 0],
[0, 0.6, 0.6], [0, -0.6, 0.6], [0, 0.6, -0.6], [0, -0.6, -0.6]])
self.assertAlmostEqual(ftmd.anisotropy(planar), 0.25, places=2)
def test_anisotropy_star(self):
star = np.array([[0., 0, 1], [0, 1, 0], [1, 0, 0],
[-1, 0, 0], [0, -1, 0], [0, 0, 1], [0, 0, 0]])
self.assertLessEqual(ftmd.anisotropy(star), 0.2)
def test_anisotropy_linear(self):
linear = np.array([[1., 1., 1.], [0., 0., 0.],
[-1., -1., -1.], [-2, -2, -2]])
self.assertAlmostEqual(ftmd.anisotropy(linear), 1)
def describe_rna(cg, file_num, dist_pais, angle_pairs):
data = {}
data["nt_length"] = cg.seq_length
data["num_cg_elems"] = len(cg.defines)
for letter in "smifth":
data["num_" + letter] = len([x for x in cg.defines if x[0] == letter])
multiloops = cg.find_mlonly_multiloops()
descriptors = []
junct3 = 0
junct4 = 0
reg = 0
pk = 0
op = 0
for ml in multiloops:
descriptors = cg.describe_multiloop(ml)
if "regular_multiloop" in descriptors:
if len(ml) == 3:
junct3 += 1
elif len(ml) == 4:
junct4 += 1
reg += 1
if "pseudoknot" in descriptors:
pk += 1
if "open" in descriptors:
op += 1
data["3-way-junctions"] = junct3
data["4-way-junctions"] = junct4
#print (descriptors)
data["open_mls"] = op
# print(data["open_mls"][-1])
data["pseudoknots"] = pk
data["regular_mls"] = reg
data["total_mls"] = len(multiloops)
try:
data["longest_ml"] = max(len(x) for x in multiloops)
except ValueError:
data["longest_ml"] = 0
try:
data["rog_fast"] = cg.radius_of_gyration("fast")
except (ftmc.RnaMissing3dError, AttributeError):
data["rog_fast"] = float("nan")
data["rog_vres"] = float("nan")
data["anisotropy_fast"] = float("nan")
data["anisotropy_vres"] = float("nan")
data["asphericity_fast"] = float("nan")
data["asphericity_vres"] = float("nan")
else:
data["rog_vres"] = cg.radius_of_gyration("vres")
data["anisotropy_fast"] = ftmd.anisotropy(cg.get_ordered_stem_poss())
data["anisotropy_vres"] = ftmd.anisotropy(
cg.get_ordered_virtual_residue_poss())
data["asphericity_fast"] = ftmd.asphericity(cg.get_ordered_stem_poss())
data["asphericity_vres"] = ftmd.asphericity(
cg.get_ordered_virtual_residue_poss())
for from_nt, to_nt in dist_pairs:
try:
dist = ftuv.vec_distance(cg.get_virtual_residue(int(from_nt), True),
cg.get_virtual_residue(int(to_nt), True))
except Exception as e:
dist = float("nan")
log.warning("%d%s File %s: Could not calculate distance between "
"%d and %d: %s occurred: %s", file_num,
{1: "st", 2: "nd", 3: "rd"}.get(
file_num % 10 * (file_num % 100 not in [11, 12, 13]), "th"),
cg.name, from_nt, to_nt, type(e).__name__, e)
data["distance_{}_{}".format(from_nt, to_nt)] = dist
for elem1, elem2 in angle_pairs:
try:
angle = ftuv.vec_angle(cg.coords.get_direction(elem1),
cg.coords.get_direction(elem2))
except Exception as e:
angle = float("nan")
log.warning("%d%s File %s: Could not calculate angle between "
"%s and %s: %s occurred: %s", file_num,
{1: "st", 2: "nd", 3: "rd"}.get(
file_num % 10 * (file_num % 100 not in [11, 12, 13]), "th"),
cg.name, elem1, elem2, type(e).__name__, e)
data["angle_{}_{}".format(elem1, elem2)] = angle
data["missing_residues_5prime"] = (len(cg.seq.with_missing[:1]) - 1)
data["missing_residues_3prime"] = (
len(cg.seq.with_missing[cg.seq_length:]) - 1)
data["missing_residues_middle"] = (
len(cg.seq.with_missing[1:cg.seq_length]) - len(cg.seq[1:cg.seq_length]))
data["missing_residues_total"] = (
len(cg.seq.with_missing[:]) - len(cg.seq[:]))
fp = len(cg.seq.with_missing[:1]) - 1
tp = 0
old_bp = None
bp = None
for bp in cg.backbone_breaks_after:
fp += len(cg.seq.with_missing[bp:bp + 1].split('&')[1]) - 1
tp += len(cg.seq.with_missing[bp:bp + 1].split('&')[0]) - 1
tp += len(cg.seq.with_missing[cg.seq_length:]) - 1
data["missing_residues_5prime_chain"] = (fp)
data["missing_residues_3prime_chain"] = (tp)
data["missing_residues_middle_chain"] = (
data["missing_residues_total"] - fp - tp)
incomplete_elem_types = Counter(x[0] for x in cg.incomplete_elements)
data["s_with_missing"] = incomplete_elem_types["s"]
data["i_with_missing"] = incomplete_elem_types["i"]
data["m_with_missing"] = incomplete_elem_types["m"]
data["h_with_missing"] = incomplete_elem_types["h"]
mp = ""
if incomplete_elem_types["s"]:
for elem in cg.incomplete_elements:
if elem[0] != "s":
continue
for i in range(cg.defines[elem][0], cg.defines[elem][1]):
left_s = cg.seq.with_missing[i:i + 1]
if len(left_s) > 2:
right_s = cg.seq.with_missing[cg.pairing_partner(
i + 1):cg.pairing_partner(i)]
if len(right_s) > 2:
mp += "{}&{};".format(left_s, right_s)
data["missing_basepairs"] = mp
return data
def anisotropy(cgs):
ai = np_nans(len(cgs))
for i, cg in enumerate(cgs):
ai[i] = ftmd.anisotropy(cg.get_ordered_stem_poss())
return ai
