sigfit

[!WARNING]
This package is in early development and not yet ready for use

The sigfit R package implements common mutational signature fitting algorithms that allow you to identify the most likely combination of known mutational signatures which explain the mutational profile observed in a sample of interest.

[!TIP]
Sigfit is exclusively a signature fitting tool. For turnkey end-to-end signature analysis we recommend sigscreen

Unlike most other implementations, sigfit expects no specific signature collection or feature space. Users supply a catalogue of counts & a matching signature collection, so any type of signature analysis / feature space is supported.

Installation

You can install the development version of sigfit like so:

if (!require("remotes"))
    install.packages("remotes")

remotes::install_github("selkamand/sigfit")

Quick Start

There are 2 inputs required.

1. observed mutation counts: A named numeric vector where names represent channels and values represent counts
2. signature collection: A channel X signature matrix where values represent fractions.

library(sigfit)

catalogue <- c(channel1 = 20, channel2 = 104, channel3 = 10)

signatures <- matrix(
  c(0, 0.9, 0.1, 0.98, 0.01, 0.01, 0.25, 0.5, 0.25),
  nrow = 3L,
  ncol = 3L,
  dimnames = list(c("channel1", "channel2", "channel3"), c("Sig1", "Sig2", "Sig3"))
)

sig_fit(catalogue, signatures, method = "nnls")