LCD: Linear Combination Decomposition
In YAPSA: Yet Another Package for Signature Analysis

Description Usage Arguments Value See Also Examples

LCD performs a mutational signatures decomposition 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

1	LCD(in_mutation_catalogue_df, in_signatures_df, in_per_sample_cutoff = 0)

`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_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.

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

lsei

## define raw data
W_prim <- matrix(c(1,2,3,4,5,6),ncol=2)
W_prim_df <- as.data.frame(W_prim)
W_df <- YAPSA:::normalize_df_per_dim(W_prim_df,2) # corresponds to the sigs
W <- as.matrix(W_df)
## 1. Simple case: non-negativity already in raw data
H <- matrix(c(2,5,3,6,1,9,1,2),ncol=4)
H_df <- as.data.frame(H) # corresponds to the exposures
V <- W %*% H # matrix multiplication
V_df <- as.data.frame(V) # corresponds to the mutational catalogue
exposures_df <- YAPSA:::LCD(V_df,W_df)
## 2. more complicated: raw data already contains negative elements
## define indices where sign is going to be swapped
sign_ind <- c(5,7)
## now compute the indices of the other fields in the columns affected
## by the sign change
row_ind <- sign_ind %% dim(H)[1]
temp_ind <- 2*row_ind -1
other_ind <- sign_ind + temp_ind
## alter the matrix H to yield a new mutational catalogue
H_compl <- H
H_compl[sign_ind] <- (-1)*H[sign_ind]
H_compl_df <- as.data.frame(H_compl) # corresponds to the exposures
V_compl <- W %*% H_compl # matrix multiplication
V_compl_df <- as.data.frame(V_compl) # corresponds to the mutational catalog
exposures_df <- YAPSA:::LCD(V_compl_df,W_df)
exposures <- as.matrix(exposures_df)