def construct_pose_from_matching_domains( host_pose, # pose
host_res1, # int
host_res2, # int
guest_name, # '1aaaA.pdb'
guest_res1, # int
guest_res2 ): # int
# build guest pose
guest_pose = Pose()
pose_from_pdb( guest_pose, 'alpha-beta-hydrolases/'+guest_name )
# rotate guest pose to align with host's ref residues
kabsch_alignment( host_pose, guest_pose, [ host_res1 - 1 ,
host_res1 ,
host_res1 + 1 ,
host_res2 - 1 ,
host_res2 ,
host_res2 + 1 ],
[ guest_res1 - 1 ,
guest_res1 ,
guest_res1 + 1 ,
guest_res2 - 1 ,
guest_res2 ,
guest_res2 + 1 ] )
pymover.apply(host_pose)
pymover.apply(guest_pose)
raw_input('see pymol for match')
# generate new pose from aligned domains
new_pose = Pose()
f.write( str(host_pose) )
f.write( str(guest_pose) )
f.flush()
def main():
opts, args = getopt.getopt(sys.argv[3:], 'i')
show_index = 0
for o in opts:
if '-i' in o:
show_index = 1
rosetta.init()
wtName = sys.argv[1]
compareName = sys.argv[2]
outputName = wtName.split('.')[0] + '_vs_' + compareName.split('.')[0] + ".txt"
pose1 = rosetta.pose_from_pdb(wtName)
pose2 = rosetta.pose_from_pdb(compareName)
use_me = True
if pose1.total_residue() != pose2.total_residue():
print "Residue number not equal", pose1.total_residue(), \
pose2.total_residue()
use_me = False
else:
output = open(outputName, 'w')
total_residue = pose1.total_residue()
kabsch_alignment(pose1, pose2, range(1, total_residue + 1), range(1, total_residue + 1))
# RMSD calculated by my own function
for i in range(1, total_residue + 1):
calculateRMS(pose1, pose2, i, output, show_index)
# RMSD calculated by PyRosetta
ro_rmsd = rosetta.CA_rmsd(pose1, pose2)
print "rosetta generated rmsd: " + str(ro_rmsd)
if use_me:
global total_square
me_rmsd = math.sqrt(total_square / total_residue)
print "me generated rmsd: " + str(me_rmsd)
output.write(outputName.split('.')[0] + ":\t" + str(ro_rmsd))
output.close()
print "Done"
def main():
opts, args = getopt.getopt(sys.argv[3:], 'i')
show_index = 0
for o in opts:
if '-i' in o:
show_index = 1
rosetta.init()
file_index = 1
wtName = sys.argv[1]
compareName = sys.argv[2]
relax_name = relax_filename(compareName, file_index)
outputName = wtName.split('.')[0] + '_vs_' + compareName + ".txt"
output = open(outputName, 'w')
pose1 = rosetta.Pose(wtName)
total_residue = pose1.total_residue()
rmsdList = [0.00] * total_residue
totalRMSD = []
while os.path.isfile(relax_name):
pose2 = rosetta.Pose(relax_name)
if pose1.total_residue() != pose2.total_residue():
print "Residue number not equal"
break
else:
kabsch_alignment(pose1, pose2, range(1, total_residue + 1), range(1, total_residue + 1))
for i in range(1, total_residue + 1):
calculateRMS(pose1, pose2, i, rmsdList)
ro_rmsd = rosetta.CA_rmsd(pose1, pose2)
totalRMSD.append(ro_rmsd)
print "rosetta generated rmsd: " + str(ro_rmsd)
global total_square
me_rmsd = math.sqrt(total_square / total_residue)
print "me generated rmsd: " + str(me_rmsd)
total_square = 0.0
file_index = file_index + 1
relax_name = relax_filename(compareName, file_index)
if file_index == 1:
print "No relaxation file"
else:
file_index = file_index - 1
rmsd_total = 0.0
if file_index > 0:
for rmsd in totalRMSD:
rmsd_total = rmsd_total + rmsd
averageRMSD = rmsd_total / file_index
print "average rmsd: ", str(averageRMSD)
outputMinMax(totalRMSD, compareName)
print "outputing " + outputName + "..."
for index in range(1, total_residue + 1):
rmsdList[index - 1] = rmsdList[index - 1] / file_index
if show_index:
output.write(str(index) + '\t' + str(rmsdList[index - 1]) + '\n')
else:
output.write(str(rmsdList[index - 1]) + '\n')
output.write(outputName.split('.')[0] + "_relax\taverage rmsd: " + str(averageRMSD))
output.close()
print "Done"
def construct_pose_from_matching_domains( old_host_pose, # pose
host_res1, # int
host_res2, # int
old_guest_pose, # pose
guest_res1, # int
guest_res2 ): # int
# copy poses
# TODO is this step actually necessary? any significant speed toll?
host_pose = Pose()
host_pose.assign( old_host_pose )
guest_pose = Pose()
guest_pose.assign( old_guest_pose )
# rotate guest pose to align with host's ref residues
kabsch_alignment( host_pose, guest_pose, [ host_res1 - 1 ,
host_res1 ,
host_res1 + 1 ],
# host_res2 - 1 ,
# host_res2 ,
# host_res2 + 1 ],
[ guest_res1 - 1 ,
guest_res1 ,
guest_res1 + 1 ])#,
# guest_res2 - 1 ,
# guest_res2 ,
# guest_res2 + 1 ] )
# define residues for new pose
host_rsds = [r for r in range(1, host_res1+1)] + [r for r in range(host_res2, host_pose.total_residue())]
guest_rsds = [r for r in range(guest_res1, guest_res2+1)]
# check clashiness
n=1
close_ones = 0
for gn in guest_rsds:
for hn in host_rsds:
dist = guest_pose.residue( gn ).xyz('CA').distance( host_pose.residue( hn ).xyz('CA') )
if dist < 3.0:
close_ones += 1
if close_ones > .01*args.clash_percentage*len(guest_rsds):
return 1
# too much trouble just to get pdb_id.pdb, annoying.
# hope that pymol chooses this as the name
host_pose_name = host_pose.pdb_info().name().split('/')[-1].split('.')[0]
guest_pose_name = guest_pose.pdb_info().name().split('/')[-1].split('.')[0]
print 'HOST: ', host_pose_name
print 'GUEST: ', guest_pose_name
print "####################"
print 'rmsd ', rmsd
print "CLASHES: ", close_ones
print 'guest_res1 ', guest_res1
print 'guest_res2 ', guest_res2
print "####################"
if args.visualize:
pymover = PyMOL_Mover()
pymover.apply(host_pose)
time.sleep(.1)
# prettify the visualization
pymol.cmd.hide('everything')
pymol.cmd.show('cartoon')
time.sleep(.1)
# easiest to see is pink cartoon for host pose and yellow lines for guest loop
pymol.cmd.color('pink')
pymover.apply(guest_pose)
time.sleep(.1)
pymol.cmd.hide('everything')
pymol.cmd.show('cartoon')
time.sleep(.1)
pymol.cmd.color('yellow', guest_pose_name)
# hide guest pose residues outside of the loop of interest
pymol.cmd.hide( '( not resi ' + str(guest_res1)+'-'+str(guest_res2) + ' and '+guest_pose_name+' )' )
# hide host pose residues outside of lame loop
pymol.cmd.hide( '( '+host_pose_name+' and resi ' + str(host_res1)+'-'+str(host_res2) + ' )' )
# show old loop slightly transparent and gray?
raw_input('hit enter for CE')
# CEALIGN
pymol.cmd.cealign(host_pose_name, guest_pose_name)
# analyze
pymol.cmd.save('ce_host.pdb', host_pose_name)
#time.sleep?
pymol.cmd.save('ce_guest.pdb', guest_pose_name)
ce_host_pose = Pose()
ce_guest_pose = Pose()
pose_from_pdb(ce_host_pose, 'ce_host.pdb')
pose_from_pdb(ce_guest_pose, 'ce_guest.pdb')
print ce_host_pose
print ce_guest_pose
if args.find_residues:
##### MIN DISTS #####
#print "* see file ce_dists.csv for output"
#all_ce_dist_mins = []
#for r1 in range(1,ce_host_pose.total_residue()+1):
# ce_dists = []
# for r2 in range(1, ce_guest_pose.total_residue()+1):
# try:
# dist = ce_host_pose.residue( r1 ).xyz('CA').distance( ce_guest_pose.residue( r2 ).xyz('CA') )
# ce_dists.append(dist)
# except:
# print "exception"
# min_ce_dist = min(ce_dists)
# all_ce_dist_mins.append(min_ce_dist)
#with open('ce_dists.csv', 'w') as dists_file:
# for item in all_ce_dist_mins:
# dists_file.write( str(item)+'\n' ) # any need for csv module?
#print "ce_dists file written"
##### SEC STRUCT #####
print "checking secondary structures to preserve"
# USING PDB NUMBERING
print "\n\nUSING PDB NUMBERING\n\n"
# populate pose secstructs
DsspMover().apply(ce_host_pose)
DsspMover().apply(ce_guest_pose)
ce_host_ss = ce_host_pose.secstruct()
ce_guest_ss = ce_guest_pose.secstruct()
print ce_host_ss
print ce_guest_ss
print "# host res1"
print 'pose_num:', host_res1
print 'pdb_num: ', PDBN(ce_host_pose,host_res1)
# find min_dist-guest-res from manually chosen host res
ce_res_dists = {}
for r1 in range(1, ce_guest_pose.total_residue()+1):
try:
dist = ce_host_pose.residue( host_res1 ).xyz('CA').distance( ce_guest_pose.residue( r1 ).xyz('CA') )
ce_res_dists[PDBN(ce_guest_pose,r1)] = dist
except:
print "exception"
closest_res = min(ce_res_dists, key=ce_res_dists.get)
print "# dists"
print [str(res)+': '+str(dist) for res, dist in ce_res_dists.iteritems() if closest_res-10 < res < closest_res+10 ]
print "# closest guest res"
print closest_res
# move to end of ss if present
#TODO confirm accuracy in indexing
for r1 in range(closest_res-1, ce_guest_pose.total_residue()): # index offset
if ce_guest_ss[r1] == 'L':
guest_res1_ss_end = PDBN(ce_guest_pose,r1)
break
print "# last res of guest ss"
print guest_res1_ss_end
print "# guest's regional ss"
for a in range(closest_res-10,closest_res+10):
print PDBN(ce_guest_pose,a+1), ce_guest_ss[a]
# closest guest ss-end upstream
# closest guest ss-end downstream
raw_input('hit enter to continue to next match')
if args.outfiles:
# already working with duplicate poses, just modify in place
# to remove residues, delete backwards; rosetta updates sequence index to keep continuity from 1
for r in reversed(range(host_res1, host_res2+1)): # WILL SEGFAULT AT r = 1
host_pose.delete_polymer_residue( r )
for r in reversed(guest_rsds):
host_pose.append_polymer_residue_after_seqpos( guest_pose.residue( r ), host_res1, 0 )
lego_pose = Pose()
lego_pose.assign( host_pose )
# TODO shorten pose name after debugged and stuff
ref_pose = host_pose_name+'-'+str(host_res1)+'-'+str(host_res2)
origin_pose = host_pose_name+'-'+str(host_res1)+'-'+str(host_res2)
match_pose = guest_pose_name+'-'+str(guest_res1)+'-'+str(guest_res2)
#match_params = '%.4f-%i' % (rmsd, close_ones)
try:
os.mkdir( 'pose-dumps/' )
except OSError:
pass # dir exists
dump_name = 'pose-dumps/'+origin_pose+'_'+match_pose+'.pdb'
lego_pose.dump_pdb( dump_name )
print 'successfully dumped to file ' + dump_name
if __name__ == '__main__':
parser = argparse.ArgumentParser()
# not optional
parser.add_argument( "ref_pdb", help="path to host pdb file" )
parser.add_argument( "ref_res_1", type=int, help="residue 1 for reference pdb" )
parser.add_argument( "ref_res_2", type=int, help="residue 2 for reference pdb" )
m_group = parser.add_mutually_exclusive_group()
# directory of guests
m_group.add_argument( "-d", "--directory", help="directory of pdbs to check against" )
# specific guest
m_group.add_argument( "-g", "--guest_pdb", help="path to a single pdb to check against" )
# specific reses in guest
parser.add_argument( "--res1", type=int, help="use only these residues in guest_pdb" )
parser.add_argument( "--res2", type=int, help="use only these residues in guest_pdb" )
# output file
parser.add_argument( "-o", "--outfiles", help="whether or not pdbs are created (in pose-dumps/ dir) from the matches", action="store_true" )
# visualize
parser.add_argument( "-V", "--visualize", help="use import-capable pymol (not MacPyMOL) to visualize matches as they are found", action="store_true" )
# tolerances
parser.add_argument( "--rmsd", default=3, help="maximum root-mean-squared deviation between transforms (default is 3)" )
parser.add_argument( "--length", default=100,help="maximum number of residues in matching loop (default is 100)" )
parser.add_argument( "--clash_percentage", default=10, help="maximum number of residues that clash between new loop and host, as a percentage of the length of the new loop (default is 10)" )
parser.add_argument( "--loop_percentage", default=50, help="maximum percentage of new loop as 'loop' secondary structure (default is 50)" )
parser.add_argument( "--open_tolerances", help="set all tolerances to maximum", action="store_true" )
# "--half-tolerances"; if passed, set all tolerances to half
# cealign distances
parser.add_argument( "--find_residues", help="guess which residues to clip at", action="store_true" )
args = parser.parse_args()
# maximize tolerances
if args.open_tolerances:
args.rmsd = 100
args.length = 10000
args.clash_percentage = 100
args.loop_percentage = 100
print 'tolerances have been set to maximum'
# start pymol if appropriate
if args.visualize:
print 'importing pymol...'
sys.path.append('/opt/local/Library/Frameworks/Python.framework/Versions/2.6/lib/python2.6/site-packages/')
import pymol
silence = pymol.finish_launching()
silence = pymol.cmd.do('run ~/Desktop/PyRosetta/PyMOLPyRosettaServer.py')
time.sleep(2)
if True: #with open( 'thorp-log', 'w') as loggy, open('matching_domains.'+str(time.strftime("%Y%m%d-%H%M%S")), 'w') as matchy:
orig_pdb_file = args.ref_pdb # ./path/to/aaaa.pdb
orig_pdb_id = orig_pdb_file.split('/')[-1] # aaaa.pdb
residue1 = args.ref_res_1
residue2 = args.ref_res_2
# DEBUG OUTPUT
#output_string = '\n\n#######################################\n'+ \
# 'This file contains debugging output for thorp.py\n'+ \
# 'invoked at: '+str(time.strftime("%Y/%m/%d %H:%M:%S"))+'\n\n\n'+ \
# 'Reference PDB: '+str(orig_pdb_file)+'\n'+ \
# 'Residues '+str(residue1)+' and '+str(residue2)+'\n' #+ \
# #str(len(os.listdir(family_dir)))+' dir items (PDBs to posify and traverse for transforms)\n\n'
#loggy.write(output_string)
#loggy.flush()
# TRANSFORMS OUTPUT
# TODO output tolerances
#matchy_string = \
# 'This list from *** directory, ### pdbs, *** tolerances, etc \n' + \
# 'saved at: '+str(time.strftime("%Y/%m/%d %H:%M:%S"))+'\n\n'+ \
# str(sys.argv)+'\n\n\n'
#matchy.write(matchy_string)
#matchy.flush()
find_matching_domains( args )
def construct_pose_from_matching_domains(
old_host_pose, # pose
host_res1, # int
host_res2, # int
old_guest_pose, # pose
guest_res1, # int
guest_res2): # int
# copy poses
# TODO is this step actually necessary? any significant speed toll?
host_pose = Pose()
host_pose.assign(old_host_pose)
guest_pose = Pose()
guest_pose.assign(old_guest_pose)
# rotate guest pose to align with host's ref residues
kabsch_alignment(
host_pose,
guest_pose,
[host_res1 - 1, host_res1, host_res1 + 1],
# host_res2 - 1 ,
# host_res2 ,
# host_res2 + 1 ],
[guest_res1 - 1, guest_res1, guest_res1 + 1]) #,
# guest_res2 - 1 ,
# guest_res2 ,
# guest_res2 + 1 ] )
# define residues for new pose
host_rsds = [r for r in range(1, host_res1 + 1)
] + [r for r in range(host_res2, host_pose.total_residue())]
guest_rsds = [r for r in range(guest_res1, guest_res2 + 1)]
# check clashiness
n = 1
close_ones = 0
for gn in guest_rsds:
for hn in host_rsds:
dist = guest_pose.residue(gn).xyz('CA').distance(
host_pose.residue(hn).xyz('CA'))
if dist < 3.0:
close_ones += 1
if close_ones > .01 * args.clash_percentage * len(guest_rsds):
return 1
# too much trouble just to get pdb_id.pdb, annoying.
# hope that pymol chooses this as the name
host_pose_name = host_pose.pdb_info().name().split('/')[-1].split('.')[0]
guest_pose_name = guest_pose.pdb_info().name().split('/')[-1].split('.')[0]
print 'HOST: ', host_pose_name
print 'GUEST: ', guest_pose_name
print "####################"
print 'rmsd ', rmsd
print "CLASHES: ", close_ones
print 'guest_res1 ', guest_res1
print 'guest_res2 ', guest_res2
print "####################"
if args.visualize:
pymover = PyMOL_Mover()
pymover.apply(host_pose)
time.sleep(.1)
# prettify the visualization
pymol.cmd.hide('everything')
pymol.cmd.show('cartoon')
time.sleep(.1)
# easiest to see is pink cartoon for host pose and yellow lines for guest loop
pymol.cmd.color('pink')
pymover.apply(guest_pose)
time.sleep(.1)
pymol.cmd.hide('everything')
pymol.cmd.show('cartoon')
time.sleep(.1)
pymol.cmd.color('yellow', guest_pose_name)
# hide guest pose residues outside of the loop of interest
pymol.cmd.hide('( not resi ' + str(guest_res1) + '-' +
str(guest_res2) + ' and ' + guest_pose_name + ' )')
# hide host pose residues outside of lame loop
pymol.cmd.hide('( ' + host_pose_name + ' and resi ' + str(host_res1) +
'-' + str(host_res2) + ' )')
# show old loop slightly transparent and gray?
raw_input('hit enter for CE')
# CEALIGN
pymol.cmd.cealign(host_pose_name, guest_pose_name)
# analyze
pymol.cmd.save('ce_host.pdb', host_pose_name)
#time.sleep?
pymol.cmd.save('ce_guest.pdb', guest_pose_name)
ce_host_pose = Pose()
ce_guest_pose = Pose()
pose_from_pdb(ce_host_pose, 'ce_host.pdb')
pose_from_pdb(ce_guest_pose, 'ce_guest.pdb')
print ce_host_pose
print ce_guest_pose
if args.find_residues:
##### MIN DISTS #####
#print "* see file ce_dists.csv for output"
#all_ce_dist_mins = []
#for r1 in range(1,ce_host_pose.total_residue()+1):
# ce_dists = []
# for r2 in range(1, ce_guest_pose.total_residue()+1):
# try:
# dist = ce_host_pose.residue( r1 ).xyz('CA').distance( ce_guest_pose.residue( r2 ).xyz('CA') )
# ce_dists.append(dist)
# except:
# print "exception"
# min_ce_dist = min(ce_dists)
# all_ce_dist_mins.append(min_ce_dist)
#with open('ce_dists.csv', 'w') as dists_file:
# for item in all_ce_dist_mins:
# dists_file.write( str(item)+'\n' ) # any need for csv module?
#print "ce_dists file written"
##### SEC STRUCT #####
print "checking secondary structures to preserve"
# USING PDB NUMBERING
print "\n\nUSING PDB NUMBERING\n\n"
# populate pose secstructs
DsspMover().apply(ce_host_pose)
DsspMover().apply(ce_guest_pose)
ce_host_ss = ce_host_pose.secstruct()
ce_guest_ss = ce_guest_pose.secstruct()
print ce_host_ss
print ce_guest_ss
print "# host res1"
print 'pose_num:', host_res1
print 'pdb_num: ', PDBN(ce_host_pose, host_res1)
# find min_dist-guest-res from manually chosen host res
ce_res_dists = {}
for r1 in range(1, ce_guest_pose.total_residue() + 1):
try:
dist = ce_host_pose.residue(host_res1).xyz('CA').distance(
ce_guest_pose.residue(r1).xyz('CA'))
ce_res_dists[PDBN(ce_guest_pose, r1)] = dist
except:
print "exception"
closest_res = min(ce_res_dists, key=ce_res_dists.get)
print "# dists"
print[
str(res) + ': ' + str(dist)
for res, dist in ce_res_dists.iteritems()
if closest_res - 10 < res < closest_res + 10
]
print "# closest guest res"
print closest_res
# move to end of ss if present
#TODO confirm accuracy in indexing
for r1 in range(closest_res - 1,
ce_guest_pose.total_residue()): # index offset
if ce_guest_ss[r1] == 'L':
guest_res1_ss_end = PDBN(ce_guest_pose, r1)
break
print "# last res of guest ss"
print guest_res1_ss_end
print "# guest's regional ss"
for a in range(closest_res - 10, closest_res + 10):
print PDBN(ce_guest_pose, a + 1), ce_guest_ss[a]
# closest guest ss-end upstream
# closest guest ss-end downstream
raw_input('hit enter to continue to next match')
if args.outfiles:
# already working with duplicate poses, just modify in place
# to remove residues, delete backwards; rosetta updates sequence index to keep continuity from 1
for r in reversed(range(host_res1,
host_res2 + 1)): # WILL SEGFAULT AT r = 1
host_pose.delete_polymer_residue(r)
for r in reversed(guest_rsds):
host_pose.append_polymer_residue_after_seqpos(
guest_pose.residue(r), host_res1, 0)
lego_pose = Pose()
lego_pose.assign(host_pose)
# TODO shorten pose name after debugged and stuff
ref_pose = host_pose_name + '-' + str(host_res1) + '-' + str(host_res2)
origin_pose = host_pose_name + '-' + str(host_res1) + '-' + str(
host_res2)
match_pose = guest_pose_name + '-' + str(guest_res1) + '-' + str(
guest_res2)
#match_params = '%.4f-%i' % (rmsd, close_ones)
try:
os.mkdir('pose-dumps/')
except OSError:
pass # dir exists
dump_name = 'pose-dumps/' + origin_pose + '_' + match_pose + '.pdb'
lego_pose.dump_pdb(dump_name)
print 'successfully dumped to file ' + dump_name
if terms[2]=='O1' and terms[3]=='TST' and terms[4]=='X':
tst_o1 = [ float(terms[6]), float(terms[7]), float(terms[8]) ]
if terms[2]=='C2' and terms[3]=='TST' and terms[4]=='X':
tst_c2 = [ float(terms[6]), float(terms[7]), float(terms[8]) ]
break
except:
pass
pose = pose_from_pdb( tst_type_set, pathname )
ap1_tmp, pose1 = structural_alignment.kabsch_alignment( ap1_tmp,
pose,
[ ap1_c33,
ap1_s1,
ap1_o9,
ap1_c29 ],
[ tst_c1,
tst_s1,
tst_o1,
tst_c2 ] )
ap1_out = None
ap1_out = Pose()
ap1_out.assign(ap1_tmp)
pose = pose_from_pdb( tst_type_set, pathname )
ap2_tmp, pose2 = structural_alignment.kabsch_alignment( ap2_tmp,
pose,
[ ap2_c1,
ap2_s1,
ap2_o1,
ap2_c3 ],
[ tst_c1,