Integrating personalized gene expression profiles into predictive disease-associated gene pools

Abstract

Gene expression data are routinely used to identify genes that on average exhibit different expression levels between a case and a control group. Yet, very few of such differentially expressed genes are detectably perturbed in individual patients. Here, we develop a framework to construct personalized perturbation profiles for individual subjects, identifying the set of genes that are significantly perturbed in each individual. This allows us to characterize the heterogeneity of the molecular manifestations of complex diseases by quantifying the expression-level similarities and differences among patients with the same phenotype. We show that despite the high heterogeneity of the individual perturbation profiles, patients with asthma, Parkinson and Huntington’s disease share a broadpool of sporadically disease-associated genes, and that individuals with statistically significant overlap with this pool have a 80–100% chance of being diagnosed with the disease. The developed framework opens up the possibility to apply gene expression data in the context of precision medicine, with important implications for biomarker identification, drug development, diagnosis and treatment.

A personalised approach for identifying disease-relevant pathways in heterogeneous diseases

“…With the increasing affordability of high-throughput technologies, such as microarray and RNA sequencing, genome-wide time-course gene expression data has become one of the most abundant and routinely analysed type of data [Bar-Joseph et al, 2012] for studying and understanding the molecular mechanisms underlying various complex diseases [Menche et al, 2017]. Encapsulating a wealth of information regarding the prolonged or transient expressions of a large set of activated genes [Bar-Joseph et al, 2012], time-course data also helps us understand and model the (multidimensional) dynamics of complex biological systems or phenomena, such as disease progression [Bar-Joseph et al, 2012;Androulakis et al, 2007;Wang et al, 2008].…”

“…The traditional applications of these methods detect genes that exhibit different expression levels between a case and a control group (DEGs) across the whole study population. Unfortunately, in the case of heterogeneous data from complex diseases, only a few genes are usually found to be DE across all or most cases because different genes with similar functionalities may be found to be perturbed across cases, thus justifying the gene-level variability at a functional or pathway level [Menche et al, 2017]. In fact, gene-level results from similar studies of heterogeneous diseases, such as cancers [Segal et al, 2004;Drier et al, 2013], asthma, Huntington's diseases [Menche et al, 2017], rheumatoid arthritis, type 2 diabetes, schizophrenia [Jin et al, 2014], and Parkinson's disease [Jin et al, 2014;Menche et al, 2017], have often been found to be inconsistent.…”

“…Unfortunately, in the case of heterogeneous data from complex diseases, only a few genes are usually found to be DE across all or most cases because different genes with similar functionalities may be found to be perturbed across cases, thus justifying the gene-level variability at a functional or pathway level [Menche et al, 2017]. In fact, gene-level results from similar studies of heterogeneous diseases, such as cancers [Segal et al, 2004;Drier et al, 2013], asthma, Huntington's diseases [Menche et al, 2017], rheumatoid arthritis, type 2 diabetes, schizophrenia [Jin et al, 2014], and Parkinson's disease [Jin et al, 2014;Menche et al, 2017], have often been found to be inconsistent. They show distressingly little overlap between similar studies of the same disease [Subramanian et al, 2005;Segal et al, 2004;Chen et al, 2013;Menche et al, 2017].…”

“…In fact, gene-level results from similar studies of heterogeneous diseases, such as cancers [Segal et al, 2004;Drier et al, 2013], asthma, Huntington's diseases [Menche et al, 2017], rheumatoid arthritis, type 2 diabetes, schizophrenia [Jin et al, 2014], and Parkinson's disease [Jin et al, 2014;Menche et al, 2017], have often been found to be inconsistent. They show distressingly little overlap between similar studies of the same disease [Subramanian et al, 2005;Segal et al, 2004;Chen et al, 2013;Menche et al, 2017]. Due to these challenges, many methods that summarise the results on a pathway-level have been developed, where the genes are unified under biological themes that aid in a functional understanding of the results.…”

“…by analysing each case-control pair separately; and a pathway is an overarching term for a group of genes unified under biological themes and are also referred to as gene sets in Subramanian et al [2005]. Menche et al [2017] introduced a framework for personalised gene expression analysis, where personalised perturbation profiles (PEEPs) are constructed per case subject by calculating a z-score with reference to the control group and considering any gene with a z-score above an optimised threshold to be part of the PEEP. Using a combinatorial model on the PEEPs, they strive to identify a single pool of disease-associated genes that can be used to accurately predict the disease status of each subject.…”