Description Usage Arguments Details Value Examples

View source: R/SpectralTAD_Par.R

Parallelized Hierarchical Spectral Clustering of TADs

1 2 3 4 |

`cont_list` |
List of contact matrices where each is in either sparse 3 column, n x n or n x (n+3) form, where the first 3 columns are chromosome, start and end coordinates of the regions. If an x n matrix is used, the column names must correspond to the start point of the corresponding bin. Required. |

`chr` |
Vector of chromosomes in the same order as their corresponding contact matrices. Must be same length as cont_list. Required. |

`levels` |
The number of levels of the TAD hierarchy to be calculated. The default setting is 1. |

`qual_filter` |
Option to turn on quality filtering which removes TADs with negative silhouette scores (poorly organized TADs). Default is FALSE. |

`z_clust` |
Option to filter sub-TADs based on the z-score of their eigenvector gaps. Default is TRUE. |

`eigenvalues` |
The number of eigenvectors to be calculated. The default and suggested setting is 2. |

`min_size` |
The minimum allowable TAD size measured in bins. Default is 5. |

`window_size` |
The size of the sliding window for calculating TADs. Smaller window sizes correspond to less noise from long-range contacts but limit the possible size of TADs |

`resolution` |
The resolution of the contact matrix. If none selected, the resolution is estimated by taking the most common distance between bins. For n x (n+3) contact matrices, this value is automatically calculated from the first 3 columns. |

`grange` |
Parameter to determine whether the result should be a GRangeList object. Defaults to FALSE |

`gap_threshold` |
Corresponds to the percentage of zeros allowed before a column/row is removed from analysis. 1=100%, .7 = 70%, etc. Default is 1. |

`cores` |
Number of cores to use. Defaults to total available cores minus one. |

`labels` |
Vector of labels used to name each contact matrix. Must be same length as cont_list. Default is NULL. |

This is the parallelized version of the SpectralTAD() function. Given a sparse 3 column, an n x n contact matrix, or n x (n+3) contact matrix, SpectralTAD returns a list of TAD coordinates in BED format. SpectralTAD works by using a sliding window that moves along the diagonal of the contact matrix. By default we use the biologically relevant maximum TAD size of 2Mb and minimum size of 5 bins to determine the size of this window. Within each window, we calculate a Laplacian matrix and determine the location of TAD boundaries based on gaps between eigenvectors calculated from this matrix. The number of TADs in a given window is calculated by finding the number that maximize the silhouette score. A hierarchy of TADs is created by iteratively applying the function to sub-TADs. The number of levels in each hierarchy is determined by the user.

List of lists where each entry is a list of data frames or GRanges in BED format corresponding to TADs seperated by hierarchies

1 2 3 4 5 6 7 8 9 | ```
#Read in data
data("rao_chr20_25_rep")
#Make a list of matrices
mat_list = list(rao_chr20_25_rep, rao_chr20_25_rep)
#Make a vector of chromosomes
chr = c("chr20", "chr20")
#Make a vector of labels
labels = c("run1", "run2")
spec_table <- SpectralTAD_Par(mat_list, chr= chr, labels = labels)
``` |

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.