MinibatchDiagnosticPlot

To ensure sufficient convergence of the loss for jrSiCKLSNMF with mini-batch updates, we
plot the loss vs the number of iterations for the mini-batch algorithm.
After a certain number of iterations, the loss should appear to oscillate
around a value. Before continuing with downstream analyses, please ensure that
the loss exhibits this sort of behavior. For the mini-batch algorithm, it is not
possible to use the convergence criteria used for the batch version of the
algorithm.

Methods to perform Joint graph Regularized Single-Cell Kullback-Leibler Sparse Non-negative Matrix Factorization ('jrSiCKLSNMF', pronounced "junior sickles NMF") on quality controlled single-cell multimodal omics count data. 'jrSiCKLSNMF' specifically deals with dual-assay scRNA-seq and scATAC-seq data. This package contains functions to extract meaningful latent factors that are shared across omics modalities. These factors enable accurate cell-type clustering and facilitate visualizations. Methods for pre-processing, clustering, and mini-batch updates and other adaptations for larger datasets are also included. For further details on the methods used in this package please see Ellis, Roy, and Datta (2023) <doi:10.3389/fgene.2023.1179439>.

Dorothy Ellis

jrSiCKLSNMF

Multimodal Single-Cell Omics Dimensionality Reduction

Susmita Datta

Kenneth Perkins

Renaud Gaujoux

MinibatchDiagnosticPlot function

<dl><dt>SickleJr</dt>
<dd>An object of class SickleJr</dd></dl>

Arguments

Plot a diagnostic plot for the mini-batch algorithm — MinibatchDiagnosticPlot

<dl>

<dt>SickleJr</dt>
<dd>An object of class SickleJr</dd>

</dl>

MinibatchDiagnosticPlot: Plot a diagnostic plot for the mini-batch algorithm

Description

Usage

Value

Arguments

Examples