def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
assert cluster.compute_similarity(
activesite_a,
activesite_b) == cluster.compute_similarity(activesite_b, activesite_a)
assert cluster.compute_similarity(activesite_a, activesite_a) == 1
assert cluster.compute_similarity(activesite_b, activesite_b) == 1
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
# update this assertion
assert cluster.compute_similarity(activesite_a,
activesite_b) == 0.959349593495935
assert cluster.compute_similarity(activesite_a, activesite_a) == 0.0
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
assert cluster.compute_similarity(activesite_a,
activesite_b) == float("inf")
assert round(cluster.compute_similarity(activesite_a, activesite_a),
1) == 0.0
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
# testing that the distance between the two sites is as expected
assert cluster.compute_similarity(activesite_a,
activesite_b) == 26.464581285329228
# testing the the distance between a site and itself is 0
assert cluster.compute_similarity(activesite_a, activesite_a) == 0
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
# update this assertion
#check that the distance between A and B is the same as between B and A
assert cluster.compute_similarity(
activesite_a,
activesite_b) == cluster.compute_similarity(activesite_b, activesite_a)
#check that the distance between A and A is 0
assert cluster.compute_similarity(activesite_a, activesite_a) == 0.0
#check that the distance is always positive
assert cluster.compute_similarity(activesite_a, activesite_a) >= 0.0
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
assert cluster.compute_similarity(
activesite_a.counts,
activesite_b.counts) == np.linalg.norm(activesite_a.counts -
activesite_b.counts)
assert cluster.compute_similarity(activesite_a.counts,
activesite_a.counts) == 0
assert cluster.compute_similarity(
activesite_a.counts,
activesite_b.counts) == cluster.compute_similarity(
activesite_b.counts, activesite_a.counts)
def test_distance(): #test that similarity between two same ones is 100%
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "276.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
# update this assertion
assert cluster.compute_similarity(activesite_a, activesite_b) == 100.0
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
# update this assertion
#First test: reflexivity of this distance metric
assert cluster.compute_similarity(activesite_a, activesite_a) == 0.0
#Second test: symmetric
trans_1 = cluster.compute_similarity(activesite_a, activesite_b)
trans_2 = cluster.compute_similarity(activesite_b, activesite_a)
assert trans_1 == trans_2
#Third test: non-negativity
assert trans_1, trans_2 >= 0
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
assert cluster.compute_similarity(activesite_a,
activesite_b,
comparator="hydrophobicity") == 11.1556
def test_similarity():
filename_a = os.path.join("data", "46495.pdb")
filename_b = os.path.join("data", "23812.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
cluster.get_order_residues([activesite_a])
cluster.get_order_residues([activesite_b])
# update this assertion
assert (cluster.compute_similarity(activesite_a, activesite_b)) == 13
def assign_single_site_to_cluster(site, centroids):
'''
check which cluster centroid is closest to the given site and assign the
site to that cluster
'''
loc = site.counts
dists = {}
for c in centroids.keys():
dist = cl.compute_similarity(loc, centroids[c])
dists[dist] = c
closest = dists[min(dists.keys())]
return closest
def test_similarity():
filename_a = os.path.join("data", "276.pdb")
filename_b = os.path.join("data", "4629.pdb")
activesite_a = io.read_active_site(filename_a)
activesite_b = io.read_active_site(filename_b)
# dist(a,b) == expected
assert round(compute_similarity(activesite_a, activesite_b), 3) == 13.857
# dist(a,a) == 0
assert compute_similarity(activesite_a, activesite_a) == 0
# dist(a,b) == dist(b,a)
assert compute_similarity(activesite_a,
activesite_b) == compute_similarity(
activesite_b, activesite_a)
# sign(dist(a,b)) == +
active_sites = read_active_sites("data")
for i in range(len(active_sites)):
for j in range(len(active_sites)):
if i != j:
assert compute_similarity(active_sites[i], active_sites[j]) > 0
def compute_similarity_matrix(sites):
''' copy of computer similarity matrix from utils '''
simMat = []
names = []
for i in range(len(sites)):
names.append(sites[i].name)
row = []
for j in range(len(sites)):
row.append(cl.compute_similarity(sites[i].counts, sites[j].counts))
simMat.append(row)
simMat = pd.DataFrame(simMat, columns=names, index=names)
return simMat
def compute_new_cluster_sim(new_clust_avg, simMat_update, sites_dict,
clusters):
'''
compute the similarity of the newly created cluster to the rest of the clusters
'''
newSim = []
for site in simMat_update.columns:
if site not in sites_dict:
s = compute_cluster_center(clusters[site], sites_dict, aa_df)
else:
s = sites_dict[site].counts
newSim.append(cl.compute_similarity(new_clust_avg, s))
newSim.append(0.0)
return newSim
def compute_similarity_matrix(sites):
'''
Compute the similarity matrix of all sites by all sites using the similarity algorithm in the cluster.py script
In: list of sites
out: similarity matrix as pandas dataframe
'''
simMat = []
names = []
for i in range(len(sites)):
names.append(sites[i].name)
row = []
for j in range(len(sites)):
row.append(cl.compute_similarity(sites[i].counts, sites[j].counts))
simMat.append(row)
simMat = pd.DataFrame(simMat, columns=names, index=names)
return simMat
def check_change_in_centroids(old_centroids, new_centroids):
''' check how far the centroids have moved '''
diff = 0
for c in old_centroids.keys():
diff += cl.compute_similarity(old_centroids[c], new_centroids[c])
return diff