def for_lysozyme():
"""
try various version of distance matrix for only lysozyme
"""
import sys
method = sys.argv[1]
paths = [
"CEpiMatch.csv",
"Multiprot.csv",
"SPa.csv",
"SPb.csv",
"SPe.csv",
"TMa.csv",
"TMb.csv",
"TMc.csv",
"MATT.csv",
"RMSD.csv",
]
for path in paths:
mats = categorize(load_mat("data/%s" % path), data)
mat = mats["lysozyme"]
if path not in ("RMSD.csv",):
mat = 1 / mat
if method == "upper":
mat = np.triu(mat)
elif method == "lower":
mat = np.transpose(np.tril(mat))
elif method == "average":
mat = (np.tril(mat) + np.triu(mat)) / 2
plt.figure()
Z = linkage(mat)
dendrogram(Z, labels=mat.rlabels, orientation="right")
plt.title("Hierarchical clustering for lysozyme")
plt.ylabel("PDB ID")
plt.savefig("img/for_lysozyme_%s/%s.png" % (method, path.split(".")[0]))
def compare():
"""
plot a series of hcluster dendragram for various methods on various sets of virus
"""
paths = [
"CEpiMatch.csv",
"Multiprot.csv",
"SPa.csv",
"SPb.csv",
"SPe.csv",
"TMa.csv",
"TMb.csv",
"TMc.csv",
"MATT.csv",
"RMSD.csv",
]
for path in paths:
mats = categorize(load_mat("data/%s" % path), data)
for name, mat in mats.items():
if path not in ("RMSD.csv",):
mat = 1 / mat
Z = linkage(mat)
plt.figure()
dendrogram(Z, labels=mat.rlabels, orientation="right")
type_name = path.split(".")[0]
plt.title("%s %s" % (name, type_name))
plt.ylabel("PDB ID")
plt.savefig("img/%s/%s.png" % (name, type_name))