Prediction of gene expression with cis-SNPs using mixed models and regularization methods

Ping Zeng,Xiang Zhou,Shuiping Huang

doi:10.1186/s12864-017-3759-6

Ping Zeng, Xiang Zhou + Show 1 more

Open Access

https://doi.org/10.1186/s12864-017-3759-6

Copy DOI

Abstract

BackgroundIt has been shown that gene expression in human tissues is heritable, thus predicting gene expression using only SNPs becomes possible. The prediction of gene expression can offer important implications on the genetic architecture of individual functional associated SNPs and further interpretations of the molecular basis underlying human diseases.MethodsWe compared three types of methods for predicting gene expression using only cis-SNPs, including the polygenic model, i.e. linear mixed model (LMM), two sparse models, i.e. Lasso and elastic net (ENET), and the hybrid of LMM and sparse model, i.e. Bayesian sparse linear mixed model (BSLMM). The three kinds of prediction methods have very different assumptions of underlying genetic architectures. These methods were evaluated using simulations under various scenarios, and were applied to the Geuvadis gene expression data.ResultsThe simulations showed that these four prediction methods (i.e. Lasso, ENET, LMM and BSLMM) behaved best when their respective modeling assumptions were satisfied, but BSLMM had a robust performance across a range of scenarios. According to R2 of these models in the Geuvadis data, the four methods performed quite similarly. We did not observe any clustering or enrichment of predictive genes (defined as genes with R2 ≥ 0.05) across the chromosomes, and also did not see there was any clear relationship between the proportion of the predictive genes and the proportion of genes in each chromosome. However, an interesting finding in the Geuvadis data was that highly predictive genes (e.g. R2 ≥ 0.30) may have sparse genetic architectures since Lasso, ENET and BSLMM outperformed LMM for these genes; and this observation was validated in another gene expression data. We further showed that the predictive genes were enriched in approximately independent LD blocks.ConclusionsGene expression can be predicted with only cis-SNPs using well-developed prediction models and these predictive genes were enriched in some approximately independent LD blocks. The prediction of gene expression can shed some light on the functional interpretation for identified SNPs in GWASs.

Highlights

It has been shown that gene expression in human tissues is heritable, predicting gene expression using only SNPs becomes possible
One way to explain this is that the identified SNPs are associated with molecular-level traits, such as methylation levels and gene expression levels, which are thought to mediate the effects of SNPs on many complex traits and diseases, and hold the key to understand the genetic basis of disease susceptibility and phenotypic variation
In this paper we explore to predict gene expression with only cis-SNPs by borrowing two risk prediction models that are well studied and widely used in genome wide association studies (GWASs), i.e. linear mixed model (LMM) [46, 54,55,56,57] and Bayesian sparse linear mixed model (BSLMM) [58]

Summary

Introduction

It has been shown that gene expression in human tissues is heritable, predicting gene expression using only SNPs becomes possible. In the last decade tens of thousands of SNPs have been identified by genome wide association studies (GWASs) for many complex human diseases and traits [1,2,3], such as type I and II diabetes [4,5,6,7], lung cancer [8,9,10,11], Crohn’s disease [12, 13], rheumatoid arthritis [13,14,15,16,17,18], blood pressure and hypertension [19,20,21], prostate cancer [22,23,24,25,26], height [27, 28], schizophrenia and bipolar disorder [29], and many others These successes offer unprecedented insights into the genetic architectures of human diseases and traits, and may lead to clinically promising preventions and treatments for diseases in the future [30, 31]. Investigation of gene expression measurements can offer important implications on the genetic architecture of individual functional associated SNPs and provide further interpretations of the molecular basis underlying human diseases [32, 35, 37, 38]

Methods

Results

Conclusion