hicPCA¶
Computes PCA eigenvectors for a Hi-C matrix.
$ hicPCA –matrix hic_matrix.h5 -o pca1.bedgraph pca2.bedgraph
usage: hicPCA --matrix MATRIX --outputFileName OUTPUTFILENAME
[OUTPUTFILENAME ...]
[--whichEigenvectors WHICHEIGENVECTORS [WHICHEIGENVECTORS ...]]
[--format {bedgraph,bigwig}]
[--chromosomes CHROMOSOMES [CHROMOSOMES ...]]
[--method {dist_norm,lieberman}] [--ligation_factor]
[--extraTrack EXTRATRACK] [--histonMarkType HISTONMARKTYPE]
[--pearsonMatrix PEARSONMATRIX] [--obsexpMatrix OBSEXPMATRIX]
[--ignoreMaskedBins] [--help] [--version]
Required arguments¶
- --matrix, -m
HiCExplorer matrix in h5 format.
- --outputFileName, -o
File names for the result of the pca.Number of output files must match the number of computed eigenvectors.
Optional arguments¶
- --whichEigenvectors, -we
The list of eigenvectors that the PCA should compute e.g. 1 2 5 will return the first, second and fifth eigenvector. (Default: “1 2”).
Default: “1 2”
- --format, -f
Possible choices: bedgraph, bigwig
Output format. Either bedgraph or bigwig (Default: “bigwig”).
Default: “bigwig”
- --chromosomes
List of chromosomes to be included in the correlation.
- --method
Possible choices: dist_norm, lieberman
possible methods which can be used to build the obs-exp matrix are dist_norm and lieberman (Default: “dist_norm”).
Default: “dist_norm”
- --ligation_factor
Setting this flag multiplies a scaling factor to each entry of the expected matrix to take care of the proximity ligation as has been explained in Homer software. This flag is only affective with dist_norm method and will be ignored if lieberman method is chosen.
Default: False
- --extraTrack
Either a gene track or a histone mark coverage file (preferably a broad mark) is needed to decide if the values of the eigenvector need a sign flip or not. Please consider: bed files are interpreted as gene tracks and bigwig files as histone marks.
- --histonMarkType
Set it to active or inactive. This is only necessary if a histon mark coverage file is given as an extraTrack (Default: “active”).
Default: “active”
- --pearsonMatrix, -pm
Internally the input matrix is converted per chromosome to obs_exp matrix and consecutively to a Pearson matrix. Set this parameter to write the pearson matrix to a file.
- --obsexpMatrix, -oem
Internally the input matrix is converted per chromosome to obs_exp matrix and consecutively to a Pearson matrix. Set this parameter to write the observe/expected matrix to a file.
- --ignoreMaskedBins
Mask bins are usually set to 0. This option removes the masked bins before the PCA is computed. Attention: this will lead to empty PCA regions.
Default: False
- --version
show program’s version number and exit