Skip to content

essHIC.ess

stefanofranzini edited this page Sep 27, 2020 · 1 revision
essHIC.ess(indir,norm,chromo,res,masked=[])

The ess class provides methods to compare each couple of HiC maps in the dataset and compute the essential distance matrix.


Parameters:

indir: string
path to the directory that contains all experiments.
norm: string
normalization of the HiC matrices to compare.
chromo: integer
number of the chromosome to compare
res: string
resolution of the HiC matrices to compare.

Attributes:

input_path: string
path of the directory containing the experiments.
norm: string
normalization of the HiC matrices to compare.
chromo: integer
chromosome of the HiC matrices to compare.
res: string
resolution of the HiC matrices to compare.
filelist: list of strings
list of paths of the matrices to compare.
nmatrix: integer
number of matrices to compare.
max_nvec: integer
number of eigenspaces to compute for each HiC matrix
eig: dictionary of numpy ndarray
dictionary of the eigenvalues of each experiment. The keys are the names of the experiments
eigv: dictionary of numpy ndarray
dictionary of the eigenvectors of each experiment. The keys are the names of the experiments
banned: dictionary of numpy ndarray
dictionary of the indices of the empty columns of each experiment. The keys are the names of the experiments
significant: dictionary of numpy ndarray
of the significant eigenspaces for each experiment. The keys are the names of the experiments.


Methods

method function
get_spectra computes spectra of all experiments.
get_pseudo_spectra computes spectra of all experiments and pseudo-replicate experiments.
test_random tests which eigenvectors are not significant.
get_essential_distance computes the essential distance matrix.

__init__(indir,norm,chromo,res,masked=[])

initialize self.


get_spectra

get_spectra(self,nvec)

computes the spectrum of each experiment, up to nvec eigenspaces.

Parameters:

nvec: integer
number of eigenvectors and eigenvalues to compute for each experiment.

Returns:

none

get_pseudo_spectra

get_pseudo_spectra(self,nvec,npseudo=-1,from_norm='nrm')

computes the spectra of the experiments and of pseudo-replicates obtained from them up to nvec eigenspaces. It also writes a pseudo metadata file.

Parameters:

nvec: integer
number of eigenvectors and eigenvalues to compute for each experiment.
npseudo: integer, default=-1
number of pseudo-replicates to create. If negative, computes pseudo-replicates until each cell type as the same number of experiments in the dataset.
from_norm: string, default='nrm'
normalization from which to sample the pseudo replicates.

Returns:

none

test_random

test_random(self,pvalue=1)

computes which eigenspaces are significant, according to the given p-value.

Parameters:

pvalue: float, default=1
p-value below which eigenspaces are significant.

Returns:

none

get_essential_distance

get_essential_distance(self,output,nvec=-1,spectrum='flat')

computes and writes a essential distance matrix which contains the distances between all couples of HiC maps in the dataset.

Parameters:

output: string
path of the output file where the essential distance matrix is written.
nvec: {'test',integer}, default=-1
number of eigenspaces to use to compute the essential distance matrix. If negative, it uses all eigenspaces computed. if test, only uses significant eigenspaces.
spectrum: {'flat','norm','none',float}, default=1.:
normalization of the eigenvalue. See the hic class for more information about the normalizations.

Returns:

none

Clone this wiki locally