Skip to contents

thin wrapper around sigminer feature tally

Source: R/cn_features.R
extract_sigminer_wang_features.Rd

Reshapes/renames our typical dataframes into one sigminer is happy with and runs sigminer::sig_tally with the "W" method.

Usage

extract_sigminer_wang_features(segs_df)

Arguments

segs_df

dataframe of CN segments.

Value

matrix of feature counts.

Details

Features are based on the paper:

Wang et al. Copy number signature analysis tool and its application in prostate cancer reveals distinct mutational processes and clinical outcomes. PLOS Genetics. 2021.

https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1009557

Some issues, I, Ben Furman, have with it are:

  • any oscillations of CN count as an oscillation, including something like 2 - 500 - 2. Oscillations are supposed to reflect potential chromothripsis, which a pattern like that is not likely to be from.

  • segment sizes are binned based on log10 values. This means there are a lot of CN size bins < 1 Mb, and DLP starts at 500Kb (and segments that small should probably be filtered anyway). Thus, many of the segment size bins are not used with DLP. This isn't a sigminer issue, but an issue with the scale of DLP data.