def test_dump():
fh = open("%s/dump_01.test.dat" % outdir,'w')
rvecs,resi = bb.dump_rvec(fname,cutoff=1.7)
stri = ""
for i1 in range(len(resi)):
for i2 in range(len(resi)):
if(sum((rvecs[0,i1,i2])**2) > 1.0e-5):
stri += "%10s %10s %14e %14e %14e \n" % (resi[i1],resi[i2],rvecs[0,i1,i2,0],rvecs[0,i1,i2,1],rvecs[0,i1,i2,2])
fh.write(stri)
fh.close()
gvecs,resi = bb.dump_gvec(fname)
fh = open("%s/dump_02.test.dat" % outdir,'w')
stri = ""
for i1 in range(len(resi)):
for i2 in range(len(resi)):
if(sum((gvecs[0,i1,i2])**2) > 1.0e-5):
stri += "%10s %10s %14e %14e %14e %14e \n" % (resi[i1],resi[i2],gvecs[0,i1,i2,0],gvecs[0,i1,i2,1],gvecs[0,i1,i2,2],gvecs[0,i1,i2,3])
fh.write(stri)
fh.close()
comp("%s/dump_01.test.dat" % refdir)
comp("%s/dump_02.test.dat" % refdir)
def test_cluster():
# first, calculate all g-vectors
print("# Calculating G-vectors")
gvec,seq = bb.dump_gvec(traj,top)
lent = gvec.shape[0]
gvec = gvec.reshape(lent,-1)[::5]
print("# Calculating PCA. gvec shape: ", gvec.shape)
# calculate PCA
v,w = cc.pca(gvec,nevecs=3)
print("# Cumulative explained variance of component: 1=%5.1f 2:=%5.1f 3=%5.1f" % (v[0]*100,v[1]*100,v[2]*100))
print("# DBSCAN clustering...")
# do DBSCAN clustering. eps and min_samples need to be adjusted.
new_labels, center_idx = cc.dbscan(gvec,range(gvec.shape[0]),eps=0.6/np.sqrt(8.),min_samples=10)
print("DONE!")
# create color palette. gray, small points for unassigned clusters.
#cp = sns.color_palette("hls",len(center_idx)+1)
#colors = [cp[j-1] if(j!=0) else (0.77,0.77,0.77) for j in new_labels]
size = [5 if(j!=0) else 0.25 for j in new_labels]
# do scatterplot
#plt.scatter(w[:,0],w[:,1],s=size,c=colors)
# now dump centroids and print labels on plot
print("# Dump PDB centroids")
t = md.load(traj, top=top)
idxs = [ii for ii,kk in enumerate(new_labels) if(kk==0)]
for i,k in enumerate(center_idx):
t[k].save_pdb("%s/cluster_%03d.test.pdb" % (outdir,i))
comp("%s/cluster_%03d.test.pdb" % (refdir,i))
#plt.text(w[k,0],w[k,1],str(i),ha='center',va='center')
idxs = [ii for ii,kk in enumerate(new_labels) if(kk==i+1)]
def test_dump():
fh = open("%s/dump_01.test.dat" % outdir, 'w')
rvecs, resi = bb.dump_rvec(fname, cutoff=1.7)
stri = ""
for i1 in range(len(resi)):
for i2 in range(len(resi)):
if (sum((rvecs[0, i1, i2])**2) > 1.0e-5):
stri += "%10s %10s %14e %14e %14e \n" % (
resi[i1], resi[i2], rvecs[0, i1, i2, 0],
rvecs[0, i1, i2, 1], rvecs[0, i1, i2, 2])
fh.write(stri)
fh.close()
gvecs, resi = bb.dump_gvec(fname)
fh = open("%s/dump_02.test.dat" % outdir, 'w')
stri = ""
for i1 in range(len(resi)):
for i2 in range(len(resi)):
if (sum((gvecs[0, i1, i2])**2) > 1.0e-5):
stri += "%10s %10s %14e %14e %14e %14e \n" % (
resi[i1], resi[i2], gvecs[0, i1, i2, 0],
gvecs[0, i1, i2, 1], gvecs[0, i1, i2, 2], gvecs[0, i1, i2,
3])
fh.write(stri)
fh.close()
comp("%s/dump_01.test.dat" % refdir)
comp("%s/dump_02.test.dat" % refdir)
def test_smm_1():
fname = "%s/test/data/UUCG.pdb" % cwd
traj = "%s/test/data/UUCG.xtc" % cwd
gvec, seq = bb.dump_gvec(traj, topology=fname)
lent = gvec.shape[0]
gvec = gvec.reshape(lent, -1)[::5]
s = smm.SMM(gvec, eps=0.5)
def test_smm_1():
fname = "%s/test/data/UUCG.pdb" % cwd
traj = "%s/test/data/UUCG.xtc" % cwd
gvec,seq = bb.dump_gvec(traj,topology=fname)
lent = gvec.shape[0]
gvec = gvec.reshape(lent,-1)[::5]
s = smm.SMM(gvec,eps=0.5)
def dump(args):
assert args.dumpR or args.dumpG, "# ERROR. choose --dumpR and/or --dumpG"
if(args.dumpR):
stri_r = "# %s \n" % (" ".join(sys.argv[:]))
stri_r += "#%15s %15s %11s %11s %11s \n" % ("RES1","RES2","x","y","z")
if(args.top==None):
for i in range(len(args.pdbs)):
rvecs,resi = bb.dump_rvec(args.pdbs[i],cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
stri_r += "# PDB %s \n" % args.pdbs[i].split("/")[-1]
stri_r += "".join([" %15s %15s %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[0,i1,i2,0],rvecs[0,i1,i2,1],rvecs[0,i1,i2,2]) for i1,i2 in idxs if(sum(rvecs[0,i1,i2]**2)> 1.E-05)])
else:
rvecs,resi = bb.dump_rvec(args.trj,topology=args.top,cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
for i in range(len(rvecs)):
stri_r += "# Frame %d \n" % i
stri_r += "".join([" %15s %15s %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[i,i1,i2,0],rvecs[i,i1,i2,1],rvecs[i,i1,i2,2]) for i1,i2 in idxs if(sum(rvecs[i,i1,i2]**2)> 1.E-05)])
fh = open(args.name + ".rvec.out",'w')
fh.write(stri_r)
fh.close()
if(args.dumpG):
stri_g = "# %s \n" % (" ".join(sys.argv[:]))
stri_g += "#%15s %15s %11s %11s %11s %11s \n" % ("RES1","RES2","G0","G1","G2","G3")
if(args.top==None):
for i in range(len(args.pdbs)):
rvecs,resi = bb.dump_gvec(args.pdbs[i],cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
stri_g += "# PDB %s \n" % args.pdbs[i].split("/")[-1]
stri_g += "".join([" %15s %15s %11.4e %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[0,i1,i2,0],rvecs[0,i1,i2,1],rvecs[0,i1,i2,2],rvecs[0,i1,i2,3]) for i1,i2 in idxs if(sum(rvecs[0,i1,i2]**2)> 1.E-05)])
else:
rvecs,resi = bb.dump_rvec(args.trj,topology=args.top,cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
for i in range(len(rvecs)):
stri_g += "# Frame %d \n" % i
stri_g += "".join([" %15s %15s %11.4e %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[i,i1,i2,0],rvecs[i,i1,i2,1],rvecs[i,i1,i2,2],rvecs[i,i1,i2,3]) for i1,i2 in idxs if(sum(rvecs[i,i1,i2]**2)> 1.E-05)])
fh = open(args.name + ".gvec.out",'w')
fh.write(stri_g)
fh.close()
def dump(args):
assert args.dumpR or args.dumpG, "# ERROR. choose --dumpR and/or --dumpR"
if(args.dumpR):
stri_r = "# %s \n" % (" ".join(sys.argv[:]))
stri_r += "#%15s %15s %11s %11s %11s \n" % ("RES1","RES2","x","y","z")
if(args.top==None):
for i in range(len(args.pdbs)):
rvecs,resi = bb.dump_rvec(args.pdbs[i],cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
stri_r += "# PDB %s \n" % args.pdbs[i].split("/")[-1]
stri_r += "".join([" %15s %15s %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[0,i1,i2,0],rvecs[0,i1,i2,1],rvecs[0,i1,i2,2]) for i1,i2 in idxs if(sum(rvecs[0,i1,i2]**2)> 1.E-05)])
else:
rvecs,resi = bb.dump_rvec(args.trj,topology=args.top,cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
for i in range(len(rvecs)):
stri_r += "# Frame %d \n" % i
stri_r += "".join([" %15s %15s %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[i,i1,i2,0],rvecs[i,i1,i2,1],rvecs[i,i1,i2,2]) for i1,i2 in idxs if(sum(rvecs[i,i1,i2]**2)> 1.E-05)])
fh = open(args.name + ".rvec.out",'w')
fh.write(stri_r)
fh.close()
if(args.dumpG):
stri_g = "# %s \n" % (" ".join(sys.argv[:]))
stri_g += "#%15s %15s %11s %11s %11s %11s \n" % ("RES1","RES2","G0","G1","G2","G3")
if(args.top==None):
for i in range(len(args.pdbs)):
rvecs,resi = bb.dump_gvec(args.pdbs[i],cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
stri_g += "# PDB %s \n" % args.pdbs[i].split("/")[-1]
stri_g += "".join([" %15s %15s %11.4e %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[0,i1,i2,0],rvecs[0,i1,i2,1],rvecs[0,i1,i2,2],rvecs[0,i1,i2,3]) for i1,i2 in idxs if(sum(rvecs[0,i1,i2]**2)> 1.E-05)])
else:
rvecs,resi = bb.dump_rvec(args.trj,topology=args.top,cutoff=args.cutoff)
idxs = its.permutations(range(len(resi)), 2)
for i in range(len(rvecs)):
stri_g += "# Frame %d \n" % i
stri_g += "".join([" %15s %15s %11.4e %11.4e %11.4e %11.4e \n" % (resi[i1],resi[i2],rvecs[i,i1,i2,0],rvecs[i,i1,i2,1],rvecs[i,i1,i2,2],rvecs[i,i1,i2,3]) for i1,i2 in idxs if(sum(rvecs[i,i1,i2]**2)> 1.E-05)])
fh = open(args.name + ".gvec.out",'w')
fh.write(stri_g)
fh.close()
def test_cluster():
# first, calculate all g-vectors
print("# Calculating G-vectors")
gvec, seq = bb.dump_gvec(traj, top)
lent = gvec.shape[0]
gvec = gvec.reshape(lent, -1)[::5]
print("# Calculating PCA. gvec shape: ", gvec.shape)
# calculate PCA
v, w = cc.pca(gvec, nevecs=3)
print(
"# Cumulative explained variance of component: 1=%5.1f 2:=%5.1f 3=%5.1f"
% (v[0] * 100, v[1] * 100, v[2] * 100))
print("# DBSCAN clustering...")
# do DBSCAN clustering. eps and min_samples need to be adjusted.
new_labels, center_idx = cc.dbscan(gvec,
range(gvec.shape[0]),
eps=0.6 / np.sqrt(8.),
min_samples=10)
print("DONE!")
# create color palette. gray, small points for unassigned clusters.
#cp = sns.color_palette("hls",len(center_idx)+1)
#colors = [cp[j-1] if(j!=0) else (0.77,0.77,0.77) for j in new_labels]
size = [5 if (j != 0) else 0.25 for j in new_labels]
# do scatterplot
#plt.scatter(w[:,0],w[:,1],s=size,c=colors)
# now dump centroids and print labels on plot
print("# Dump PDB centroids")
t = md.load(traj, top=top)
idxs = [ii for ii, kk in enumerate(new_labels) if (kk == 0)]
for i, k in enumerate(center_idx):
t[k].save_pdb("%s/cluster_%03d.test.pdb" % (outdir, i))
comp("%s/cluster_%03d.test.pdb" % (refdir, i))
#plt.text(w[k,0],w[k,1],str(i),ha='center',va='center')
idxs = [ii for ii, kk in enumerate(new_labels) if (kk == i + 1)]