BASiCS

Single-cell mRNA sequencing can uncover novel cell-to-cell heterogeneity in gene expression levels within seemingly homogeneous populations of cells. However, these experiments are prone to high levels of technical noise, creating new challenges for identifying genes that show genuine heterogeneous expression within the group of cells under study.

BASiCS (Bayesian Analysis of Single-Cell Sequencing data) is an integrated Bayesian hierarchical model that propagates statistical uncertainty by simultaneously performing data normalisation (global scaling), technical noise quantification and two types of supervised downstream analyses:

• For a single group of cells [1]: BASiCS provides a criterion to identify highly (and lowly) variable genes within the group.

• For two (or more) groups of cells [2]: BASiCS allows the identification of differentially expressed genes between the groups. As in traditional differential expression tools, BASiCS can uncover changes in mean expression between the groups. Besides this, BASiCS can also uncover changes in over-dispersion --- a measure for the excess cell-to-cell variation that is observed after accounting for technical noise. This feature has led, for example, to novel insights in the context of immune cells across aging [3]. More recently, the BASiCS model has been extended to address the confounding between mean and variability that is typically observed in scRNA-seq datasets. This is achieved by introducing a residual over-dispersion parameter that is not confounded by mean expression [4].

In both cases, a probabilistic output is provided, with posterior probability thresholds calibrated through the expected false discovery rate (EFDR) [5].

The original implementation of BASiCS relies on the use of spike-in genes --- that are artificially introduced to each cell's lysate --- to perform these analyses. However, our latest work has extended the BASiCS model to datasets in which spike-ins are not available (multiple batches are required) [4].

Important: BASiCS has been designed in the context of supervised experiments where the groups of cells (e.g. experimental conditions, cell types) under study are known a priori (e.g. case-control studies). Therefore, we DO NOT advise the use of BASiCS in unsupervised settings where the aim is to uncover sub-populations of cells through clustering.

For technical details, references are provided at the bottom of this document.

Installation

BASiCS is available in Bioconductor. To install the current release use:

if (!requireNamespace("BiocManager", quietly=TRUE))
    install.packages("BiocManager")
BiocManager::install("BASiCS")

Repeat using the devel version of Bioconductor for the latest development version.

Alternatively, the experimental version of BASiCS (this might be unstable) can be installed from GitHub:

# install.packages("devtools")
devtools::install_github("catavallejos/BASiCS", build_vignettes = TRUE)

This installation might fail if some of the dependency libraries are not yet installed. If so, please run the following lines and repeat the installation.

#library(devtools)
#if (!requireNamespace("BiocManager", quietly=TRUE))
    #install.packages("BiocManager")
#BiocManager::install("BiocGenerics")
#BiocManager::install("scran")
#install.packages("Rcpp")

After a successful installation, the BASiCS library can be loaded using¹

library(BASiCS)

Installation troubleshooting

A summary of the installation errors that have been reported for BASiCS is provided here. If you encounter any additional issues, please let us know so that we can update this information.

How to use BASiCS?

BASiCS includes a vignette where its usage is illutrated. To access the vignette, please use:

vignette('BASiCS')

Individual topics are summarized in the BASiCS wiki:

• Quick start

• Input preparation

• Running the MCMC

• MCMC convergence

• Posterior summary

• HVL & LVG detection for a single group of cells

• Differential analysis between 2 groups of cells (mean and over-dispersion)

Authors

• Catalina Vallejos (cvallejos 'at' turing.ac.uk)
• Nils Eling
• Alan O'Callaghan
• John Marioni
• Sylvia Richardson

Acknowledgements

We thank several members of the Marioni laboratory (EMBL-EBI; CRUK-CI) for support and discussions throughout the development of this R library. In particular, we are grateful to Aaron Lun (@LTLA, CRUK-CI) for advise and support during the preparation the Bioconductor submission.

We also acknowledge feedback and/or contributions from (Github aliases provided within parenthesis): Alan O'Callaghan (@Alanocallaghan), Ben Dulken (@bdulken), Chang Xu (@xuchang116), Danilo Horta (@Horta), Dmitriy Zhukov (@dvzhukov), Jens Preußner (@jenzopr), Joanna Dreux (@Joannacodes), Kevin Rue-Albrecht (@kevinrue), Luke Zappia (@lazappi), Mike Morgan (@MikeDMorgan), Muad Abd El Hay (@Cumol), Nitesh Turaga (@nturaga), Simon Anders (@s-andrews), Yongchao Ge and Yuan Cao (@yuancao90), among others.

This work has been funded by the MRC Biostatistics Unit (MRC grant no. MRC_MC_UP_0801/1; Catalina Vallejos and Sylvia Richardson), EMBL European Bioinformatics Institute (core European Molecular Biology Laboratory funding; Catalina Vallejos, Nils Eling and John Marioni), CRUK Cambridge Institute (core CRUK funding; John Marioni), The Alan Turing Institute (EPSRC grant no. EP/N510129/1; Catalina Vallejos) and the University of Edinburgh (Catalina Vallejos and Alan O'Callaghan).

References

• [1] Vallejos et al. (2015). PLoS Computational Biology.
• [2] Vallejos et al. (2016). Genome Biology.
• [3] Martinez-Jimenes et al. (2017). Science.
• [4] Eling et al. (2018). Cell Systems.
• [5] Newton et al. (2004). Biostatistics.

Footnotes

1. The warning "No methods found in "BiocGenerics"" might appear. Please ignore. BASiCS imports some of the generic functions provided by BiocGenerics that do not have any methods attached. ↩