Description Usage Arguments Details Value Examples

Hierarchical Spectral Clustering of TADs

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 | ```
SpectralTAD(
cont_mat,
chr,
levels = 1,
qual_filter = FALSE,
z_clust = FALSE,
eigenvalues = 2,
min_size = 5,
window_size = 25,
resolution = "auto",
gap_threshold = 1,
grange = FALSE,
out_format = "none",
out_path = chr
)
``` |

`cont_mat` |
Contact matrix in either sparse 3 column, n x n or n x (n+3) form where the first three columns are coordinates in BED format. If an x n matrix is used, the column names must correspond to the start point of the corresponding bin. If large mode is selected, then this matrix must be a tab-seperated n x n or n x (n+3) and it should be the path to a contact matrix. Required. |

`chr` |
The chromosome of the contact matrix being analyzed. 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 three columns. |

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

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

`out_format` |
Specifies the format of the file which SpectralTAD outputs. If "none, no file is output. "juicebox" or "bedpe" returns a bedpe file compatible with juicebox. "hicexplorer" or "bed" returns a bed file compatible with hicexplorer. Default is none |

`out_path` |
Path of output file. Default is the chromosome |

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 maximizes 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.

A list where each entry is in BED format corresponding to the level of the hierarchy.

1 2 3 4 | ```
#Read in data
data("rao_chr20_25_rep")
#Find TADs
spec_table <- SpectralTAD(rao_chr20_25_rep, chr= 'chr20')
``` |

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.