Introduction of PrediXcan

• Genome-wide Association Study (GWAS) is an important tool for detecting genes related to complex diseases or biological traits, such as diabetes and cancer, animal and plant traits. Genome-wide association studies identify single nucleotide polymorphisms (which occur more frequently in individuals with diseases or target traits) to determine the mechanism of action of these associations. However, due to gene regulation, most disease-related mutations do not change the function of a gene, but change the number of genes replicated in the cell. Therefore, only through genome-wide association studies still cannot solve the problem.
• Transcriptome research can study gene expression and regulation mechanisms and their relationship with diseases to overcome this limitation of genome-wide association studies. However, transcriptome research cannot determine reverse causality-such as whether the expression level of a gene will change due to the occurrence of a disease, or whether the occurrence of a disease is caused by a change in gene expression.
• Based on the above problems, PrediXcan was developed to detect the relationship between genes and traits. PrediXcan integrates transcriptome data and GWAS data into a single computing framework. To determine the correlation between gene expression levels and disease states or target characteristics.

Fig 1. Introduction of PrediXcan.

Basic Principles of PrediXcan

The premise of PrediXcan is that the expression level of genes in the organism is mainly regulated by genetic factors, disease states and some other factors. The purpose of PrediXcan is to establish the relationship between genetically regulated gene expression and traits.

Fig 2. Mechanism tested by the PrediXcan method.

The analysis process mainly includes the following two steps:

• With the help of machine learning, statistical machine learning approaches such as Random Forest or OmicKriging are used to train the genotype data and gene expression data in GTEx Project, GEUVADIS and DGN databases, and then estimate the missing expression data in the imported genotype data. So as to estimate the gene expression level regulated by SNP.
• Based on the above data, the correlation between gene expression level and disease state or target characteristics is analyzed.

Fig 3. PrediXcan frameworks, the workflow illustrates the steps used in developing the PrediXcan method. (Eric, R, G, et al. 2015)

Advantages of PrediXcan

• Identify potential pathogenic genes
• Study the relationship between gene expression levels (high or low) and diseases or traits
• Validation studies only need to test thousands of genes at most, instead of millions of potential single-gene mutations, thereby saving computing resources
• The existing genomic data can be re-analyzed to fill the gaps in GWAS research.

Application Field

• 

  POST-GWAS ANALYSIS

• 

  TWAS ANALYSIS

• 

  eQTL ANALYSIS