perform hierarchical clustering of samples based on the mutation spectra.
Usage
cluster_spectra(
mf_data = mf_data,
group_col = "sample",
response_col = "proportion_min",
subtype_col = "normalized_subtype",
dist = "cosine",
cluster_method = "ward.D"
)Arguments
- mf_data
A data frame containing the mutation data. This data must include a column containing the mutation subtypes, a column containing the sample/cohort names, and a column containing the response variable.
- group_col
The name of the column in data that contains the sample/cohort names.
- response_col
The name of the column in data that contains the response variable. Typical response variables can be the subtype mf, proportion, or count.
- subtype_col
The name of the column in data that contains the mutation subtypes.
- dist
the distance measure to be used. This must be one of "cosine", "euclidean", "maximum", "manhattan","canberra", "binary" or "minkowski". See dist for details.
- cluster_method
The agglomeration method to be used. See hclust for details.
Details
The cosine distance measure represents the inverted cosine similarity between samples:
\(\text{Cosine Dissimilarity} = 1 - \frac{\mathbf{A} \cdot \mathbf{B}}{\| \mathbf{A} \| \cdot \| \mathbf{B} \|}\)
This equation calculates the cosine dissimilarity between two vectors A and B.
Leaves are sorted using dendsort, if installed, otherwise leaves are unsorted.