A Bayesian model for single cell transcript expression analysis on MERFISH data

Johannes Köster; Myles Brown; X Shirley Liu

doi:10.1093/bioinformatics/bty718

A Bayesian model for single cell transcript expression analysis on MERFISH data

Bioinformatics. 2019 Mar 15;35(6):995-1001. doi: 10.1093/bioinformatics/bty718.

Authors

Johannes Köster^{1

2

3}, Myles Brown^{2

4

5}, X Shirley Liu^{4

6

7}

Affiliations

¹ Algorithms for Reproducible Bioinformatics, Genome Informatics, Institute of Human Genetics, University Hospital Essen, University of Duisburg-Essen, Essen, Germany.
² Division of Molecular and Cellular Oncology, Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA.
³ Centrum Wiskunde and Informatica, Amsterdam, The Netherlands.
⁴ Center for Functional Cancer Epigenetics, Dana-Farber Cancer Institute, Boston, MA, USA.
⁵ Department of Medicine, Brigham and Women's Hospital and Harvard Medical School, Boston, MA, USA.
⁶ Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute, Harvard T.H. Chan School of Public Health, Boston, MA, USA.
⁷ School of Life Science and Technology, Tongji University, Shanghai, China.

Abstract

Motivation: Multiplexed error-robust fluorescence in-situ hybridization (MERFISH) is a recent technology to obtain spatially resolved gene or transcript expression profiles in single cells for hundreds to thousands of genes in parallel. So far, no statistical framework to analyze MERFISH data is available.

Results: We present a Bayesian model for single cell transcript expression analysis on MERFISH data. We show that the model successfully captures uncertainty in MERFISH data and eliminates systematic biases that can occur in raw RNA molecule counts obtained with MERFISH. Our model accurately estimates transcript expression and additionally provides the full probability distribution and credible intervals for each transcript. We further show how this enables MERFISH to scale towards the whole genome while being able to control the uncertainty in obtained results.

Availability and implementation: The presented model is implemented on top of Rust-Bio (Köster, 2016) and available open-source as MERFISHtools (https://merfishtools.github.io). It can be easily installed via Bioconda (Grüning et al., 2018). The entire analysis performed in this paper is provided as a fully reproducible Snakemake (Köster and Rahmann, 2012) workflow via Zenodo (https://doi.org/10.5281/zenodo.752340).

Supplementary information: Supplementary data are available at Bioinformatics online.

Publication types

Research Support, N.I.H., Extramural

MeSH terms

Bayes Theorem
Gene Expression Profiling*
In Situ Hybridization, Fluorescence
Single-Cell Analysis*
Transcription, Genetic

Abstract

Publication types

MeSH terms

Grants and funding