def scatter(rep_num, h5_path_regx):
track = loadTrack(rep_num, h5_path_regx)
ener, mcc = [], []
for i in track:
ener.append(i[1])
mcc.append(i[2])
plotScatter(mcc, ener, 'linear_scatter_ener_mcc.pdf')
def pearsonrMccEner(rep_num, h5_path_regx):
"""
read the hdf5 file and calculate the pearsonr correlation coefficient between mcc and energy
"""
track = loadTrack(rep_num, h5_path_regx)
ener, mcc = [], []
for i in track:
ener.append(i[1])
mcc.append(i[2])
return pearsonr(mcc, ener)
def extractTrack(complx, track_path):
"""
remove the duplicated data points in raw hdf5 file
store the extracted data in numpy matrix format
"""
h5_fn = complx + ".h5"
data_group = track_path
dir_path = os.path.abspath(complx)
report_path = dir_path + "/report"
h5_path_regx = dir_path + "/out*/*.h5"
num_prt, num_temp, num_lig, num_rep, temps = readReport(report_path)
print complx
for prt_conf in range(num_prt):
for lig_conf in range(num_lig):
for tmp_conf in range(num_temp):
rep_num = num_temp * num_lig * prt_conf + num_lig * tmp_conf + lig_conf
temp = temps[tmp_conf]
track = readH5.loadTrack(rep_num, h5_path_regx)
step = np.array([each_step[0] for each_step in track])
total_energy = np.array([each_step[1] for each_step in track])
mcc = np.array([each_step[2] for each_step in track])
move_vector = np.array([each_step[3] for each_step in track])
components = np.array([each_step[4] for each_step in track])
confs = np.vstack(np.array([prt_conf, lig_conf]) for i in range(len(step)))
dt = np.column_stack((step, mcc, total_energy, components, move_vector, confs))
f = h5py.File(h5_fn)
group_path = data_group
if group_path in f:
dset = f[group_path]
dset.create_dataset(str(rep_num), data=dt)
dset.attrs["prt_conf"] = prt_conf
dset.attrs["lig_conf"] = lig_conf
else:
f.create_group(group_path)
dset = f[group_path]
dset.attrs["prt_conf"] = prt_conf
dset.attrs["lig_conf"] = lig_conf
dset.create_dataset(str(rep_num), data=dt)
f.flush()
f.close()
def pearsonrMccEner(rep_num, h5_path_regx, intercept=1, floor_mcc=0.35, ceil_mcc=0.65):
"""
read the hdf5 file and calculate the pearsonr correlation coefficient between mcc and energy
"""
track = loadTrack(rep_num, h5_path_regx)
if intercept == 1:
intercepted_track = interceptTrack(floor_mcc, ceil_mcc, track)
track = intercepted_track
ener, mcc = [], []
for i in track:
ener.append(i[1])
mcc.append(i[2])
if 10000 in mcc: # drop the mcc with CMCC_INVALID_VAL
print "CMCC_INVALID_VAL found in", rep_num
return np.nan, np.nan
else:
return pearsonr(mcc, ener)
def extractMccTotalEner(complx):
"""
extract mcc and total energy from the docking tracing
store the data in hdf5 file
"""
h5_fn = complx + ".h5"
data_group = '/mcc_total'
dir_path = os.path.abspath(complx)
report_path = dir_path + '/report'
h5_path_regx = dir_path + "/out*/*.h5"
num_prt, num_temp, num_lig, num_rep, temps = readReport(report_path)
print complx
for prt_conf in range(num_prt):
for lig_conf in range(num_lig):
for tmp_conf in range(num_temp):
rep_num = num_temp * num_lig * prt_conf + num_lig * tmp_conf + lig_conf
temp = temps[tmp_conf]
track = loadTrack(rep_num, h5_path_regx)
total_energy = np.array([each_step[1] for each_step in track])
mcc = np.array([each_step[2] for each_step in track])
mcc_ener = np.column_stack((mcc, total_energy))
f = h5py.File(h5_fn)
group_path = complx + data_group
if group_path in f:
dset = f[group_path]
dset.create_dataset(str(rep_num), data=mcc_ener)
else:
f.create_group(group_path)
dset = f[group_path]
dset.create_dataset(str(rep_num), data=mcc_ener)
f.flush()
f.close()
if intercept == 1:
intercepted_track = interceptTrack(floor_mcc, ceil_mcc, track)
track = intercepted_track
ener, mcc = [], []
for i in track:
ener.append(i[1])
mcc.append(i[2])
if 10000 in mcc: # drop the mcc with CMCC_INVALID_VAL
print "CMCC_INVALID_VAL found in", rep_num
return np.nan, np.nan
else:
return pearsonr(mcc, ener)
if __name__ == "__main__":
rep_num = 362
h5_path_regx = "/work/jaydy/working/docking/1a07C1/out*/*.h5"
track = loadTrack(rep_num, h5_path_regx)
ener, mcc = [], []
for i in track:
ener.append(i[1])
mcc.append(i[2])
print pearsonr(mcc, ener)
