def test_junction_virtual_atom_distance_minimalMultiloop(self):
distance1 = ftug.junction_virtual_atom_distance(
self.minimal_multiloop, "m0")
self.assertLess(
distance1, 4., msg="{} is not < {} for {}".format(distance1, 4., "m0"))
distance2 = ftug.junction_virtual_atom_distance(
self.minimal_multiloop, "m1")
self.assertLess(
distance2, 4., msg="{} is not < {} for {}".format(distance2, 4., "m1"))
def test_junction_virtual_atom_distance_minimalMultiloop(self):
distance1 = ftug.junction_virtual_atom_distance(
self.minimal_multiloop, "m0")
self.assertLess(distance1,
4.,
msg="{} is not < {} for {}".format(
distance1, 4., "m0"))
distance2 = ftug.junction_virtual_atom_distance(
self.minimal_multiloop, "m1")
self.assertLess(distance2,
4.,
msg="{} is not < {} for {}".format(
distance2, 4., "m1"))
def test_junction_virtual_atom_distance_realPDB(self):
distance = ftug.junction_virtual_atom_distance(self.rs_random_281,
"m3")
self.assertLess(distance, 4.) #This should always hold!
self.assertAlmostEqual(
distance, 3.4294889373610675
) #This value might change, if we change the virtual atom calculation
def bulge_virtual_atom_distance(bg, ld):
'''
Calculate the distance between the O3' atom and the P
atom of the two strands that need to be closed.
@param bg: The BulgeGraph data structure
@param ld: The name of the bulge region
@return: The distance between the O3' and P' virtual atoms
of the flanking residues
'''
if len(bg.edges[ld]) == 2:
dist2 = cgg.junction_virtual_atom_distance(bg, ld)
else:
dist2 = 0.
return dist2
def describe_ml_segments(cg):
data = defaultdict(list)
loops = cg.find_mlonly_multiloops()
for loop in it.chain(loops, [[i] for i in cg.iloop_iterator()]):
print(loop)
if loop[0][0] == "i":
description = ["interior_loop"]
else:
description = cg.describe_multiloop(loop)
try:
j3_roles = cg._assign_loop_roles(loop)
except ValueError:
j3_roles = None
if j3_roles:
j3_familyFlat = cg._junction_family_westhof1(j3_roles)
j3_family3D = cg._junction_family_3d(j3_roles)
j3_familyPerp = cg._junction_family_is_perpenticular(j3_roles)
j3_Delta = cg.get_length(j3_roles["J23"]) - cg.get_length(
j3_roles["J31"])
else:
j3_family3D = None
j3_familyFlat = None
j3_familyPerp = None
j3_Delta = None
loop_start = float("inf")
for segment in loop:
if cg.define_a(segment)[0] < loop_start:
loop_start = cg.define_a(segment)[0]
for segment in loop:
if segment[0] not in "mi":
continue
data["loop_start_after"].append(loop_start)
data["segment_start_after"].append(cg.define_a(segment)[0])
data["segment"].append(segment)
data["junction_length"].append(len(loop))
data["segment_length"].append(cg.get_length(segment))
if segment[0] == "i":
dims = list(sorted(cg.get_bulge_dimensions(segment)))
else:
dims = [-1, -1]
data["iloop_length_1"].append(dims[0])
data["iloop_length_2"].append(dims[1])
data["loops_largest_segment_length"].append(
max(cg.get_length(x) for x in loop))
data["loops_shortest_segment_length"].append(
min(cg.get_length(x) for x in loop))
data["sum_of_loops_segment_lengths"].append(
sum(cg.get_length(x) for x in loop))
data["loop_segment_lengths"].append(",".join(
map(str, sorted(cg.get_length(x) for x in loop))))
data["angle_type"].append(
abs(cg.get_angle_type(segment, allow_broken=True)))
s1, s2 = cg.connections(segment)
vec1 = cg.coords.get_direction(s1)
if cg.get_sides(s1, segment) == (1, 0):
vec1 = -vec1
else:
assert cg.get_sides(s1, segment) == (0, 1)
vec2 = cg.coords.get_direction(s2)
if cg.get_sides(s2, segment) == (1, 0):
vec2 = -vec2
else:
assert cg.get_sides(s2, segment) == (0, 1)
data["angle_between_stems"].append(ftuv.vec_angle(vec1, vec2))
data["offset1"].append(
ftuv.point_line_distance(
cg.coords[s1][cg.get_sides(s1, segment)[0]],
cg.coords[s2][0], cg.coords.get_direction(s2)))
data["offset2"].append(
ftuv.point_line_distance(
cg.coords[s2][cg.get_sides(s2, segment)[0]],
cg.coords[s1][0], cg.coords.get_direction(s1)))
closer1, far1 = cg.coords[s1][cg.get_sides(
s1, segment)[0]], cg.coords[s1][cg.get_sides(s1, segment)[1]]
closer2, far2 = cg.coords[s2][cg.get_sides(
s2, segment)[0]], cg.coords[s2][cg.get_sides(s2, segment)[1]]
data["offset"].append(
ftuv.vec_distance(*ftuv.line_segment_distance(
closer1, closer1 +
(closer1 - far1) * 100000, closer2, closer2 +
(closer2 - far2) * 100000)))
data["junction_va_distance"].append(
ftug.junction_virtual_atom_distance(cg, segment))
data["is_external_multiloop"].append("open" in description)
data["is_pseudoknotted_multiloop"].append(
"pseudoknot" in description)
data["is_regular_multiloop"].append(
"regular_multiloop" in description)
data["is_interior_loop"].append("interior_loop" in description)
if j3_roles is not None:
elem_role, = [
x[0] for x in j3_roles.items() if x[1] == segment
]
else:
elem_role = "?"
data["j3_role"].append(elem_role)
data["j3_familyFlat"].append(j3_familyFlat)
data["j3_family3D"].append(j3_family3D)
data["j3_familyPerp"].append(j3_familyPerp)
data["j3_Delta_j23_j31"].append(j3_Delta)
dssr_stacking = False
if "dssr_stacks" in cg.infos:
if segment in cg.infos["dssr_stacks"]:
dssr_stacking = True
data["dssr_stacking"].append(dssr_stacking)
kh_stem_angle = float("nan")
if abs(cg.get_angle_type(segment, allow_broken=True)) == 5:
next_ml = cg.get_next_ml_segment(segment)
if isinstance(next_ml, str) and next_ml[0] == "m" and abs(
cg.get_angle_type(next_ml, allow_broken=True)) == 5:
stems1 = cg.edges[segment]
stems2 = cg.edges[next_ml]
try:
s1, s2 = (stems1 | stems2) - (stems1 & stems2)
except ValueError:
pass
else:
vec1 = cg.coords.get_direction(s1)
vec2 = cg.coords.get_direction(s2)
angle = ftuv.vec_angle(vec1, vec2)
if angle > math.pi / 2:
angle = math.pi - angle
kh_stem_angle = angle
data["kh_stem_angle"].append(kh_stem_angle)
if data:
data["pk_number"] = number_by(data, "loop_start_after",
"is_pseudoknotted_multiloop")
data["loop_number"] = number_by(data, "loop_start_after", None)
data["reguler_multiloop_number"] = number_by(data, "loop_start_after",
"is_regular_multiloop")
return data
sm.sample_stats()
possible_sstats=sm.conf_stats.sample_stats(cg, "s1")
for sstat in possible_sstats:
if sstat.pdb_name.startswith("ideal"):
break
maxdist=0
sm.elem_defs["s1"] = sstat
possible_stats=sm.conf_stats.sample_stats(cg, "m0")
for i, stat in enumerate(possible_stats):
#if i%1000 == 0:
# print(i, "/", len(possible_stats), end="")
sm.elem_defs["m0"] = stat
sm.traverse_and_build()
e=energy.eval_energy(sm)
dist = ftug.junction_virtual_atom_distance(cg, "m0")
if dist>maxdist:
maxdist=dist
if e>0:
print ("\nSampled stats do not match energy:")
print (stat)
print( dist, ">", cutoff_distance)
print(sm.elem_defs["s0"])
print(sm.elem_defs["s1"])
print( "Maxdist", maxdist, "<", cutoff_distance)
def describe_ml_segments(cg):
data = defaultdict(list)
loops = cg.find_mlonly_multiloops()
for loop in loops:
description = cg.describe_multiloop(loop)
try:
j3_roles = cg.assign_loop_roles(loop)
except ValueError:
j3_roles = None
if j3_roles:
j3_familyFlat = cg.junction_family_westhof1(j3_roles)
j3_family3D = cg.junction_family_3d(j3_roles)
j3_familyPerp = cg.junction_family_is_perpenticular(j3_roles)
j3_Delta = cg.get_length(j3_roles["J23"]) - cg.get_length(
j3_roles["J31"])
else:
j3_family3D = None
j3_familyFlat = None
j3_familyPerp = None
j3_Delta = None
loop_start = float("inf")
for segment in loop:
if cg.define_a(segment)[0] < loop_start:
loop_start = cg.define_a(segment)[0]
for segment in loop:
if segment[0] != "m":
continue
data["loop_start_after"].append(loop_start)
data["segment_start_after"].append(cg.define_a(segment)[0])
data["segment"].append(segment)
data["junction_length"].append(len(loop))
data["segment_length"].append(cg.get_length(segment))
data["loops_largest_segment_length"].append(
max(cg.get_length(x) for x in loop))
data["loops_shortest_segment_length"].append(
min(cg.get_length(x) for x in loop))
data["sum_of_loops_segment_lengths"].append(
sum(cg.get_length(x) for x in loop))
data["loop_segment_lengths"].append(",".join(
map(str, sorted(cg.get_length(x) for x in loop))))
data["angle_type"].append(
abs(cg.get_angle_type(segment, allow_broken=True)))
s1, s2 = cg.connections(segment)
vec1 = cg.coords.get_direction(s1)
if cg.get_sides(s1, segment) == (1, 0):
vec1 = -vec1
else:
assert cg.get_sides(s1, segment) == (0, 1)
vec2 = cg.coords.get_direction(s2)
if cg.get_sides(s2, segment) == (1, 0):
vec2 = -vec2
else:
assert cg.get_sides(s2, segment) == (0, 1)
data["angle_between_stems"].append(ftuv.vec_angle(vec1, vec2))
data["junction_va_distance"].append(
ftug.junction_virtual_atom_distance(cg, segment))
data["is_external_multiloop"].append("open" in description)
data["is_pseudoknotted_multiloop"].append(
"pseudoknot" in description)
data["is_regular_multiloop"].append(
"regular_multiloop" in description)
if j3_roles is not None:
elem_role, = [
x[0] for x in j3_roles.items() if x[1] == segment
]
else:
elem_role = "?"
data["j3_role"].append(elem_role)
data["j3_familyFlat"].append(j3_familyFlat)
data["j3_family3D"].append(j3_family3D)
data["j3_familyPerp"].append(j3_familyPerp)
data["j3_Delta_j23_j31"].append(j3_Delta)
if data:
data["pk_number"] = number_by(data, "loop_start_after",
"is_pseudoknotted_multiloop")
data["loop_number"] = number_by(data, "loop_start_after", None)
data["reguler_multiloop_number"] = number_by(data, "loop_start_after",
"is_regular_multiloop")
return data
def test_junction_virtual_atom_distance_realPDB(self):
distance = ftug.junction_virtual_atom_distance(
self.rs_random_281, "m3")
self.assertLess(distance, 4.) # This should always hold!
# This value might change, if we change the virtual atom calculation
self.assertAlmostEqual(distance, 3.486261134885911)
def test_junction_virtual_atom_distance_realPDB(self):
distance=ftug.junction_virtual_atom_distance(self.rs_random_281, "m4")
self.assertLess(distance, 4.)
def describe_ml_segments(cg):
data = defaultdict(list)
loops = cg.find_mlonly_multiloops()
for loop in it.chain(loops, [[i] for i in cg.iloop_iterator()]):
print(loop)
if loop[0][0] == "i":
description = ["interior_loop"]
else:
description = cg.describe_multiloop(loop)
try:
j3_roles = cg._assign_loop_roles(loop)
except ValueError:
j3_roles = None
if j3_roles:
j3_familyFlat = cg._junction_family_westhof1(j3_roles)
j3_family3D = cg._junction_family_3d(j3_roles)
j3_familyPerp = cg._junction_family_is_perpenticular(j3_roles)
j3_Delta = cg.get_length(
j3_roles["J23"]) - cg.get_length(j3_roles["J31"])
else:
j3_family3D = None
j3_familyFlat = None
j3_familyPerp = None
j3_Delta = None
loop_start = float("inf")
for segment in loop:
if cg.define_a(segment)[0] < loop_start:
loop_start = cg.define_a(segment)[0]
for segment in loop:
if segment[0] not in "mi":
continue
data["loop_start_after"].append(loop_start)
data["segment_start_after"].append(cg.define_a(segment)[0])
data["segment"].append(segment)
data["junction_length"].append(len(loop))
data["segment_length"].append(cg.get_length(segment))
if segment[0] == "i":
dims = list(sorted(cg.get_bulge_dimensions(segment)))
else:
dims = [-1, -1]
data["iloop_length_1"].append(dims[0])
data["iloop_length_2"].append(dims[1])
data["loops_largest_segment_length"].append(
max(cg.get_length(x) for x in loop))
data["loops_shortest_segment_length"].append(
min(cg.get_length(x) for x in loop))
data["sum_of_loops_segment_lengths"].append(
sum(cg.get_length(x) for x in loop))
data["loop_segment_lengths"].append(
",".join(map(str, sorted(cg.get_length(x) for x in loop))))
data["angle_type"].append(
abs(cg.get_angle_type(segment, allow_broken=True)))
s1, s2 = cg.connections(segment)
vec1 = cg.coords.get_direction(s1)
if cg.get_sides(s1, segment) == (1, 0):
vec1 = -vec1
else:
assert cg.get_sides(s1, segment) == (0, 1)
vec2 = cg.coords.get_direction(s2)
if cg.get_sides(s2, segment) == (1, 0):
vec2 = -vec2
else:
assert cg.get_sides(s2, segment) == (0, 1)
data["angle_between_stems"].append(ftuv.vec_angle(vec1, vec2))
data["offset1"].append(ftuv.point_line_distance(cg.coords[s1][cg.get_sides(s1, segment)[0]],
cg.coords[s2][0], cg.coords.get_direction(
s2)
))
data["offset2"].append(ftuv.point_line_distance(cg.coords[s2][cg.get_sides(s2, segment)[0]],
cg.coords[s1][0], cg.coords.get_direction(
s1)
))
closer1, far1 = cg.coords[s1][cg.get_sides(
s1, segment)[0]], cg.coords[s1][cg.get_sides(s1, segment)[1]]
closer2, far2 = cg.coords[s2][cg.get_sides(
s2, segment)[0]], cg.coords[s2][cg.get_sides(s2, segment)[1]]
data["offset"].append(ftuv.vec_distance(*ftuv.line_segment_distance(closer1, closer1 + (closer1 - far1) * 100000,
closer2, closer2 + (closer2 - far2) * 100000)))
data["junction_va_distance"].append(
ftug.junction_virtual_atom_distance(cg, segment))
data["is_external_multiloop"].append("open" in description)
data["is_pseudoknotted_multiloop"].append(
"pseudoknot" in description)
data["is_regular_multiloop"].append(
"regular_multiloop" in description)
data["is_interior_loop"].append("interior_loop" in description)
if j3_roles is not None:
elem_role, = [x[0]
for x in j3_roles.items() if x[1] == segment]
else:
elem_role = "?"
data["j3_role"].append(elem_role)
data["j3_familyFlat"].append(j3_familyFlat)
data["j3_family3D"].append(j3_family3D)
data["j3_familyPerp"].append(j3_familyPerp)
data["j3_Delta_j23_j31"].append(j3_Delta)
dssr_stacking = False
if "dssr_stacks" in cg.infos:
if segment in cg.infos["dssr_stacks"]:
dssr_stacking = True
data["dssr_stacking"].append(dssr_stacking)
kh_stem_angle = float("nan")
if abs(cg.get_angle_type(segment, allow_broken=True)) == 5:
next_ml = cg.get_next_ml_segment(segment)
if isinstance(next_ml, str) and next_ml[0] == "m" and abs(cg.get_angle_type(next_ml, allow_broken=True)) == 5:
stems1 = cg.edges[segment]
stems2 = cg.edges[next_ml]
try:
s1, s2 = (stems1 | stems2) - (stems1 & stems2)
except ValueError:
pass
else:
vec1 = cg.coords.get_direction(s1)
vec2 = cg.coords.get_direction(s2)
angle = ftuv.vec_angle(vec1, vec2)
if angle > math.pi / 2:
angle = math.pi - angle
kh_stem_angle = angle
data["kh_stem_angle"].append(kh_stem_angle)
if data:
data["pk_number"] = number_by(data, "loop_start_after",
"is_pseudoknotted_multiloop")
data["loop_number"] = number_by(data, "loop_start_after", None)
data["reguler_multiloop_number"] = number_by(data, "loop_start_after",
"is_regular_multiloop")
return data
def test_junction_virtual_atom_distance_realPDB(self):
distance=ftug.junction_virtual_atom_distance(self.rs_random_281, "m4")
self.assertLess(distance, 4.)
self.assertAlmostEqual(distance, 3.4765561271716967)
if __name__ == "__main__":
next_id = defaultdict(int)
args = parser.parse_args()
cgs = fuc.cgs_from_args(args, "only_cg")
for cg in cgs:
if sys.stderr.isatty():
print(cg.name, file=sys.stderr)
for elem in cg.defines.keys():
if elem in cg.incomplete_elements:
print("Skipping element", elem, file=sys.stderr)
continue
base_name = "{}:{}_".format(cg.name, elem[0])
for stat in cg.get_stats(elem):
idnr = next_id[base_name]
next_id[base_name] += 1
name = base_name + str(idnr)
stat.pdb_name = name
if elem.startswith("m"):
try:
dist = ftug.junction_virtual_atom_distance(cg, elem)
stat_dist = stat.get_virtual_atom_distance()
except:
print(stat)
#raise
else:
print(
stat, "# distance: {}. stat_dist {}".format(
dist, stat_dist))
else:
print(stat)