A multi-omics data simulator for complex disease studies and its application to evaluate multi-omics data analysis methods for disease classification

Abstract

An integrative multi-omics analysis approach that combines multiple types of omics data including genomics, epigenomics, transcriptomics, proteomics, metabolomics, and microbiomics, has become increasing popular for understanding the pathophysiology of complex diseases. Although many multi-omics analysis methods have been developed for complex disease studies, there is no simulation tool that simulates multiple types of omics data and models their relationships with disease status. Without such a tool, it is difficult to evaluate the multi-omics analysis methods on the same scale and to estimate the sample size or power when planning a new multi-omics disease study. We developed a multi-omics data simulator OmicsSIMLA, which simulates genomics (i.e., SNPs and copy number variations), epigenomics (i.e., whole-genome bisulphite sequencing), transcriptomics (i.e., RNA-seq), and proteomics (i.e., normalized reverse phase protein array) data at the whole-genome level. Furthermore, the relationships between different types of omics data, such as meQTLs (SNPs influencing methylation), eQTLs (SNPs influencing gene expression), and eQTM (methylation influencing gene expression), were modeled. More importantly, the relationships between these multi-omics data and the disease status were modeled as well. We used OmicsSIMLA to simulate a multi-omics dataset for breast cancer under a hypothetical disease model, and used the data to compare the performance among existing multi-omics analysis methods in terms of disease classification accuracy and run time. Our results 3 demonstrated that complex disease mechanisms can be simulated by OmicsSIMLA, and a random forest-based method showed the highest prediction accuracy when the multi-omics data were properly normalized.

Network and Systems Medicine: Position Paper of the European Collaboration on Science and Technology Action on Open Multiscale Systems Medicine

“…Different alternatives exist in terms of validation strategy: In silico data generation: As an example, a multi-omics data simulator for complex disease studies was developed and applied to evaluate multi-omics data analysis methods for disease classification. 106 Another tool, iOmicsPASS, allowing network-based integration of multi-omics data for predictive subnetwork discovery was recently published. 107 Validation protocols and the interpretation of validation studies: In contrast to replication, validation in other samples does not require sampling from the same populations as the discovery study.…”

“…In silico data generation: As an example, a multi-omics data simulator for complex disease studies was developed and applied to evaluate multi-omics data analysis methods for disease classification. 106 Another tool, iOmicsPASS, allowing network-based integration of multi-omics data for predictive subnetwork discovery was recently published. 107 …”

Contributions from the 2019 Literature on Bioinformatics and Translational Informatics

“…A rapid content analysis of the 931 retrieved references revealed a large proportion of studies dealing with identification and routine use in clinical settings of genetic variants in connection with various diseases. Through their choices, section editors wanted to shed light on three research trends and one emerging topic in the BTI field which are presented in the following [3][4][5][6][7][8][9][10][11][12][13][14][15][16][17]. The research trends and emerging topics cover three different dimensions: methods, application domain, and purpose.…”

“…• Early stage detection of cancer (e.g., using circulating biomarkers) and recurrence detection [16,17] ; • Patient stratification and drug sensitivity prediction, and simulation models [7,9,10,14].…”

“…This approach highlighted a significant sequence of key functions occurring at the considered stages of the disease. A quite odd paper is the one authored by Chung et al, [7] where a tool, called OmicsSIMLA, is proposed to simulate genomics (SNPs and CNV), epigenomics (i.e., bisulfite sequencing), transcriptomics, and proteomics data at the whole-genome level. Both the relationships between different types of omics data and between multi-omics data and disease status have been modeled.…”