CellMapper: rapid and accurate inference of gene expression in difficult-to-isolate cell types.

Bradlee D Nelms,Martha L Bulyk,Stuart H Orkin,Andrew W Weflen,Robert K Montgomery,Levi Waldron,Wayne I Lencer,David T Breault,Guoji Guo,Marian R Neutra,Curtis Huttenhower,Jeremy A Goettel,Luis A Barrera,Scott B Snapper

doi:10.1186/s13059-016-1062-5

Abstract

We present a sensitive approach to predict genes expressed selectively in specific cell types, by searching publicly available expression data for genes with a similar expression profile to known cell-specific markers. Our method, CellMapper, strongly outperforms previous computational algorithms to predict cell type-specific expression, especially for rare and difficult-to-isolate cell types. Furthermore, CellMapper makes accurate predictions for human brain cell types that have never been isolated, and can be rapidly applied to diverse cell types from many tissues. We demonstrate a clinically relevant application to prioritize candidate genes in disease susceptibility loci identified by GWAS.Electronic supplementary materialThe online version of this article (doi:10.1186/s13059-016-1062-5) contains supplementary material, which is available to authorized users.

Highlights

Measuring gene expression in specific cellular subsets is key to understanding cellular function and differentiation and how these processes are disrupted during disease pathogenesis
CellMapper takes as input (1) a large set of gene expression data and (2) a query gene expressed in the cell type of interest and estimates the probability that every other gene in the dataset is co-expressed with the query gene (Fig. 1a)
CellMapper is designed to make accurate predictions using as little as a single query gene, which can be selected from standard cell-specific markers employed by experimental techniques such as flow cytometry, immunohistochemistry, and promoter-driven conditional mouse knock out models

Summary

Introduction

Measuring gene expression in specific cellular subsets is key to understanding cellular function and differentiation and how these processes are disrupted during disease pathogenesis. One promising solution has been the development of computational methods to infer cell type-specific expression information directly from heterogeneous samples [8,9,10,11,12,13,14,15,16,17,18,19], such as undissociated tissue. Several machinelearning algorithms have been developed to address this problem [17,18,19], each aimed at identifying genes with a similar expression profile to an established set of cell type-specific markers, referred to here as “query genes.”

Objectives

Methods

Results

Conclusion