Introduction:
Complex traits are associated with a number of genetic variations through gene expression. There are studies that help in identifying genes associated with these traits. Genome-wide association study (GWAS) is a type of study that helps identify the genes that are associated with a particular disease. This method involves studying the whole genome (entire DNA, deoxyribonucleic acid, present in an individual’s cell) of a number of people, which helps identify variations called single nucleotide polymorphisms (SNPs). These frequently occurring variations (SNPs) help identify the genes involved in the disease development and help determine the risk of getting the disease and response to treatment approaches. Studies that involve the association of genetic variations and complex traits are transcriptome-wide association study (TWAS), colocalization, and Mendelian randomization. Transcriptome-wide association study is a locus-based method and a multi-marker association approach.
What Is Transcriptome Wide Association Studies (TWAS)?
Transcriptome-wide associations are gene-based association approaches that investigate the association between variations of phenotype and genetically regulated gene expressions. It is based on a hypothesis that one or more eQTLs (expression quantitative trait loci, which is a locus that explains the genetic variations of the gene expression) control the gene’s transcriptional activities. It reduces multiple testing and is a powerful approach to identifying genes associated with complex traits. Transcriptome-wide association studies use the phenotype (physical characteristics) data and genotype (set of genes) data from the GWAS and transcription data.
What Are The Steps Involved In Transcriptome-Wide Association Studies?
It is a two-step analysis that involves:
-
First Step: Imputation (inferring unobserved genotypes among the individuals) of gene expression by using transcriptional regulatory effects of the eQTLs for a gene under additive genetic effect. This is done in multiple tissues, separately or jointly. There are many eQTLs models, such as Genotype-Tissue Expression (GTEx), BLUEPRINT, MESA, and eQTLGEN. This helps in gathering genetic components of gene expression.
The genetic expression regulated by a gene is expressed as,
E=XW
Where X (matrix of genotypes of eQTLs) is N × M.
N is the size of the sample of a cohort study.
M is the number of eQTLs present in a gene.
W is eQTLs regulatory effects on the gene, which is taken from the transcriptome data panel (independent reference).
-
Second Step: The gene expressions which are imputed from the first step are aggregated with the phenotype of the disease of interest to estimate statistics of each disease-gene association. This step helps in the evaluation of the regression coefficient of each gene’s phenotype.
What Are The Advantages Of Transcriptome-Wide Association Studies?
The advantages of transcriptome-wide association studies are:
-
It has a high accuracy for locally heritable genes.
-
It helps in understanding the mechanism of disease.
-
Steps involved in this study can be done independently.
-
Multiple testing is eliminated in this study. This study only includes adjusting the number of tested genes which is around twenty-thousand genes. This advantage helped in the identification of the PALMD gene and calcific aortic valve stenosis.
-
These studies are tissue-specific, thus having the advantage of predicting gene expression levels from a specific tissue.
What Are The Influential Factors Of Transcriptome-Wide Association Studies?
There are three influential factors, which include:
1. The Nature Of Input GWAS Data: There are many forms of data. The PrediXcan method involves individual-level variants of eQTLs, but individual genotype data are not easily available from GWAS; it provides an accurate estimation of gene trait associations. Methods that use GWAS summary statistics are S-PrediXcan, UTMOST, and FUSION, which imputes the regression statistics between the trait and gene expression obtained directly from the GWAS summary statistics.
The advantages of GWAS summary statistics-based TWAS are;
- Than the individual-level database, the GWAS database is more efficient and can analyze the larger dataset.
- It is easy to examine the structure of linkage disequilibrium (a parameter that shows the degree of genetic variation of the allele with nearby genetic variants within a population) among eQTLs by using a reference set.
- This method can prioritize genes.
The disadvantages of GWAS summary statistics-based TWAS are;
- The difference in actual LD (linkage disequilibrium) structure and reference LD matrix can produce false results.
- This method uses a large sample size to achieve satisfactory results. Therefore this needs careful interpretation and additional checking.
2. eQTL Models: The eQTL database helps in predicting gene expression accuracy. Good quality genotype data and transcriptome data help in detecting gene expression’s genetic components and can identify eQTL in small and complex regions of genes. Genotype-Tissue Expression (GTEx) project is a well-known tissue-specific eQTL study.
3. The Methods Used To Estimate The Association Of Gene-Trait: Not only eQTLs, integrative cross-tissue analysis is also important in TWAS. There are two methods:
a. Tissue-specific TWAS: PrediXcan (tissue-specific) design is a type of tissue-specific TWAS, which can investigate the association of gene traits; it shows the statistically significant association between predicted gene expression levels and the disease of interest tissue by tissue.
Disadvantages of tissue-specific TWAS are:
- It has less number of samples, which restricts the identification of eQTLs; therefore TWAS cannot predict gene expression.
- Casual tissues of complex diseases can be unclear and therefore making special tissue difficult to carry out TWAS.
- When casual tissues are unclear, exploratory TWAS has to be done. Therefore, it gets involved in multiple testing burdens.
- Shared genetic composition of gene expression regulation and cis-eQTL can cause TWAS not to identify disease-related tissue from other tissues with the same gene expression levels.
b. Cross-tissue TWAS: It can overcome the limitations of tissue-specific TWAS. It can be used in the gene expression of complex diseases. MultiXcan design is a type of Cross tissue TWAS. This can identify the gene-level association between simulated and natural data.
Conclusion:
Understanding the genetic architecture of complex diseases is ongoing research. Transcriptome-Wide Associations studies investigate the association between variations of phenotype and genetically regulated gene expressions. It is a two-step analysis, and the Nature Of Input GWAS Data, eQTL Models, and methods of association of gene traits are important in TWAS.
