nzimmermann
/
projet_court_nz
1
0
Fork 0
Code Releases Activity

modified code (__init__) to save distance matrixes in Conformations instances

master
nicolas-zimmermann 2019-09-18 16:01:27 +02:00
parent d77136b7f4
commit dfe4789af2
1 changed files with 29 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

38
src/projet8.py
View File

@ -21,11 +21,23 @@ class Conformations:
def __init__(self, filename): def __init__(self, filename):
""" """
df : pd.DataFrame object filename : a .pdb file or several pdb file containing the conformations
Attribute :
- df : pd.DataFrame object
- filename : a list of str, paths to the structure files
- simple_dist : matrix distance computed with the identity method
Initiality empty, must be computed
- dissimilarity_dist : matrix distance computed with the
dissimilarity method, initialy empty because
the computation time is long
Each row of the dataframe is a conformation and each column a position Each row of the dataframe is a conformation and each column a position
of the sequence. of the sequence.
""" """
self.df = pd.DataFrame() self.df = pd.DataFrame()
self.filename = filename
self.identity_dist = None
self.dissimilarity_dist = None
for chain_name, chain in pbx.chains_from_files([filename]): for chain_name, chain in pbx.chains_from_files([filename]):
dihedrals = chain.get_phi_psi_angles() dihedrals = chain.get_phi_psi_angles()
pb_seq = pbx.assign(dihedrals) pb_seq = pbx.assign(dihedrals)
@ -34,7 +46,7 @@ class Conformations:
def identity(self): def identity(self):
""" """
Returns a matrix of the distance between all conformations based on Computes a distance matrix between all conformations based on
wether the PB at each position is identical or not. wether the PB at each position is identical or not.
Returns the matrix as a numpy.ndarray object. Returns the matrix as a numpy.ndarray object.
""" """
@ -43,8 +55,10 @@ class Conformations:
'o':15,'p':16} # dict to transform PB to int, 'o':15,'p':16} # dict to transform PB to int,
# necessary for the pdsit function # necessary for the pdsit function
dfnum = self.df.replace(dict_str_to_float) dfnum = self.df.replace(dict_str_to_float)
dist = squareform(pdist(dfnum, metric ='jaccard'))
self.identity_dist = dist
return squareform(pdist(dfnum, metric ='jaccard')) return dist
def dissimilarity(self, matrix=None): def dissimilarity(self, matrix=None):
""" """
@ -69,9 +83,9 @@ class Conformations:
dissimilarity[i][j] += matrix[confs.df.loc[i,k]][confs.df.loc[j,k]] dissimilarity[i][j] += matrix[confs.df.loc[i,k]][confs.df.loc[j,k]]
dissimilarity = dissimilarity + dissimilarity.T # fills the whole matrix dissimilarity = dissimilarity + dissimilarity.T # fills the whole matrix
return dissimilarity return self.dist_from_dissimilarity(dissimilarity)
def dist_from_dissimilarity(diss_matrix): def dist_from_dissimilarity(self, diss_matrix):
""" """
Using the substitution matrix from the PBxplore package, the obtained Using the substitution matrix from the PBxplore package, the obtained
dissimilarity matrix has both positive and negative values. Low value dissimilarity matrix has both positive and negative values. Low value
@ -85,14 +99,20 @@ class Conformations:
diss_matrix = -diss_matrix diss_matrix = -diss_matrix
diss_matrix = (diss_matrix - np.min(diss_matrix))/np.ptp(diss_matrix) diss_matrix = (diss_matrix - np.min(diss_matrix))/np.ptp(diss_matrix)
dist = diss_matrix * abs((np.identity(confs.df.shape[0])-1)) dist = diss_matrix * abs((np.identity(confs.df.shape[0])-1))
self.dissimilarity_dist = dist
return dist return dist
#def small_kmedoid(self, ncluster):
# pass
if __name__ == "__main__": if __name__ == "__main__":
if len(sys.argv) != 2: if len(sys.argv) < 2:
sys.exit("Error : usage '$ python3 projet8 md.pdb'") sys.exit("Error : usage '$ python3 projet8 md.pdb'")
confs = Conformations(sys.argv[1]) confs = Conformations(sys.argv[1])
print(confs.df) print(confs.df)