relu_transform: Performs the ReLU transform on mutational count data

In antoine186/convSig: Mutational Process Inference via Expectation Maximisation

Description

Performs the ReLU transform on mutational count data

Usage

relu_transform(mut_obj, five = FALSE, K = 5, iter_num = 5000)

Arguments

mut_obj	An object of class 'Shallowres' as produced by the function mut_count()
five	A boolean that specifies whether you want to apply a transformation based on a 5-nucleotide convolution window.
K	An integer that indicates the number of mutational processes you want to detect in your mutational count data via the transformation.

Value

A Feature matrix (feat), which contains the convolution weights associated with each mutational processes. An M matrix (mat), which contains the probability for all mutation types. A P matrix (P), which contains the mutational intensity/activity of each mutational process. A LOSS variable (LOSS), which displays the LOSS value achieved by the ReLU optimisation. A testing LOSS variable (test_LOSS), which displays the LOSS value achieved on the testing samples.

Examples

relu_res <- relu_transform(EMu_prepped, five = TRUE, K = 6)

antoine186/convSig documentation built on Jan. 17, 2020, 4:09 a.m.