ReadAndAnalyzeExposures: Assess how well inferred exposures match input exposures

In WuyangFF95/SynSigEval: Evaluate Results of Mutational Signature Analysis Software on Synthetic Spectra Created by Package SynSigGen

Assess how well inferred exposures match input exposures

Description

We assume that in many cases attribution programs will be run outside of R on file inputs and will generate fill outputs.

Usage

ReadAndAnalyzeExposures(
  extracted.sigs,
  reference.sigs,
  inferred.exp.path,
  ground.truth.exposures
)

Arguments

extracted.sigs	Path to file containing the extracted signature profiles.
reference.sigs	File containing signature profiles from which the synthetic data were generated.
inferred.exp.path	File containing mutation counts (exposures) of synthetic tumors which are inferred to extracted or input signatures.
ground.truth.exposures	File containing the exposures from which the synthetic catalogs were generated. This file is used to restrict assessment of signature exposures to only those signatures in reference.sigs that were actually represented in the exposures.

Details

Generates output files by calling TP_FP_FN_avg_sim

Value

A data.frame recording:

Ground.truth.exposure: sum of ground truth exposures of all tumors to all reference signatures.

Inferred.exposure: sum of inferred exposures of all tumors to all reference signatures. Here, inferred exposure of a tumor to a ground-truth signature equals to the sum of the exposures of this tumor to all extracted signatures which are most similar to a reference signature. If there is no extracted signature resembling an ground-truth signature, the inferred exposure of this ground-truth signature will be 0.

Absolute.difference: sum of absolute difference between ground-truth exposure and inferred exposure of all tumors to all reference signatures.

WuyangFF95/SynSigEval documentation built on Sept. 18, 2022, 11:41 a.m.