tss_predictor: Train TSS activity predictor

In CshlSiepelLab/DENR: Deconvolution of Expression for Nascent RNA Sequencing Data

Description

A function that trains and returns model to predict TSS activity using the same convolutional architecture as the pre-trained predictor that ships with this package. Just here in case a user wants to try training the model on alternative data/labels.

Usage

tss_predictor(train_features, train_labels, train = TRUE)

Arguments

train_features	a 3-dimensional numeric array where the dimensions are [transcript, bin, strand]. (IMPORTANT: bin must be >= 25)
train_labels	a 0/1 encoded vector of labels with 1 per transcript
train	logical. Whether to train model

Value

a trained keras model

CshlSiepelLab/DENR documentation built on July 16, 2021, 10:42 p.m.