BatChef

Like all gene expression data, single-cell data suffers from batch effects and other unwanted variations that makes accurate biological interpretations difficult. The scMerge method leverages factor analysis, stably expressed genes (SEGs) and (pseudo-) replicates to remove unwanted variations and merge multiple single-cell data. This package contains all the necessary functions in the scMerge pipeline, including the identification of SEGs, replication-identification methods, and merging of single-cell data.

BUSseq R package fits an interpretable Bayesian hierarchical model---the Batch Effects Correction with Unknown Subtypes for scRNA seq Data (BUSseq)---to correct batch effects in the presence of unknown cell types. BUSseq is able to simultaneously correct batch effects, clusters cell types, and takes care of the count data nature, the overdispersion, the dropout events, and the cell-specific sequencing depth of scRNA-seq data. After correcting the batch effects with BUSseq, the corrected value can be used for downstream analysis as if all cells were sequenced in a single batch. BUSseq can integrate read count matrices obtained from different scRNA-seq platforms and allow cell types to be measured in some but not all of the batches as long as the experimental design fulfills the conditions listed in our manuscript.

A model designed for dimensionality reduction and batch effect removal for scRNA-seq data. It is designed to be massively parallelizable using shared objects that prevent memory duplication, and it can be used with different mini-batch approaches in order to reduce time consumption. It assumes a negative binomial distribution for the data with a dispersion parameter that can be both commonwise across gene both genewise.

Correspondence analysis (CA) is a matrix factorization method, and is similar to principal components analysis (PCA). Whereas PCA is designed for application to continuous, approximately normally distributed data, CA is appropriate for non-negative, count-based data that are in the same additive scale. The corral package implements CA for dimensionality reduction of a single matrix of single-cell data, as well as a multi-table adaptation of CA that leverages data-optimized scaling to align data generated from different sequencing platforms by projecting into a shared latent space. corral utilizes sparse matrices and a fast implementation of SVD, and can be called directly on Bioconductor objects (e.g., SingleCellExperiment) for easy pipeline integration. The package also includes additional options, including variations of CA to address overdispersion in count data (e.g., Freeman-Tukey chi-squared residual), as well as the option to apply CA-style processing to continuous data (e.g., proteomic TOF intensities) with the Hellinger distance adaptation of CA.

Linnorm is an algorithm for normalizing and transforming RNA-seq, single cell RNA-seq, ChIP-seq count data or any large scale count data. It has been independently reviewed by Tian et al. on Nature Methods (https://doi.org/10.1038/s41592-019-0425-8). Linnorm can work with raw count, CPM, RPKM, FPKM and TPM.

Implements a variety of methods for batch correction of single-cell (RNA sequencing) data. This includes methods based on detecting mutually nearest neighbors, as well as several efficient variants of linear regression of the log-expression values. Functions are also provided to perform global rescaling to remove differences in depth between batches, and to perform a principal components analysis that is robust to differences in the numbers of cells across batches.