LCD_complex_cutoff: LCD with a signature-specific cutoff on exposures
In eilslabs/YAPSA: Yet Another Package for Signature Analysis

Description Usage Arguments Value See Also Examples

LCD_cutoff performs a mutational signatures decomposition by Linear Combination Decomposition (LCD) of a given mutational catalogue V with known signatures W by solving the minimization problem min(||W*H - V||) with additional constraints of non-negativity on H where W and V are known, but excludes signatures with an overall contribution less than a given signature-specific cutoff (and thereby accounting for a background model) over the whole cohort.

LCD_complex_cutoff_perPID is a wrapper for LCD_complex_cutoff and runs individually for every PID.

LCD_complex_cutoff(in_mutation_catalogue_df, in_signatures_df,
  in_cutoff_vector = NULL, in_filename = NULL, in_method = "abs",
  in_per_sample_cutoff = 0, in_rescale = TRUE, in_sig_ind_df = NULL,
  in_cat_list = NULL)

LCD_complex_cutoff_perPID(in_mutation_catalogue_df, in_signatures_df,
  in_cutoff_vector = NULL, in_filename = NULL, in_method = "abs",
  in_rescale = TRUE, in_sig_ind_df = NULL, in_cat_list = NULL)

`in_mutation_catalogue_df`	A numeric data frame `V` with `n` rows and `m` columns, `n` being the number of features and `m` being the number of samples
`in_signatures_df`	A numeric data frame `W` with `n` rows and `l` columns, `n` being the number of features and `l` being the number of signatures
`in_cutoff_vector`	A numeric vector of values less than 1. Signatures from within `W` with an overall exposure less than the respective value in `in_cutoff_vector` will be discarded.
`in_filename`	A path to generate a histogram of the signature exposures if non-NULL
`in_method`	Indicate to which data the cutoff shall be applied: absolute exposures, relative exposures
`in_per_sample_cutoff`	A numeric value less than 1. Signatures from within `W` with an exposure per sample less than `in_cutoff` will be discarded.
`in_rescale`	Boolean, if TRUE (default) the exposures are rescaled such that colSums over exposures match colSums over mutational catalogue
`in_sig_ind_df`	Data frame of type signature_indices_df, i.e. indicating name, function and meta-information of the signatures. Default is NULL.
`in_cat_list`	List of categories for aggregation. Have to be among the column names of `in_sig_ind_df`. Default is NULL.

A list with entries:

exposures: The exposures H, a numeric data frame with l rows and m columns, l being the number of signatures and m being the number of samples
norm_exposures: The normalized exposures H, a numeric data frame with l rows and m columns, l being the number of signatures and m being the number of samples
signatures: The reduced signatures that have exposures bigger than in_cutoff
choice: Index vector of the reduced signatures in the input signatures
order: Order vector of the signatures by exposure
residual_catalogue: Numerical data frame (matrix) of the difference between fit (product of signatures and exposures) and input mutational catalogue
rss: Residual sum of squares (i.e. sum of squares of the residual catalogue)
cosDist_fit_orig_per_matrix: Cosine distance between the fit (product of signatures and exposures) and input mutational catalogue computed after putting the matrix into vector format (i.e. one scaler product for the whole matrix)
cosDist_fit_orig_per_col: Cosine distance between the fit (product of signatures and exposures) and input mutational catalogue computed per column (i.e. per sample, i.e. as many scaler products as there are samples in the cohort)
sum_ind: Decreasing order of mutational loads based on the input mutational catalogue
out_sig_ind: Data frame of the type signature_indices_df, i.e. indicating name, function and meta-information of the signatures. Default is NULL, non-NULL only if in_sig_ind_df is non-NULL.
aggregate_exposures_list: List of exposure data frames aggregated over different categories. Default is NULL, non-NULL only if in_sig_ind_df and in_cat_list are non-NULL and if the categories specified in in_cat_list are among the column names of in_sig_ind_df.