A novel method for multiple phenotype association studies based on genotype and phenotype network

Abstract

Joint analysis of multiple correlated phenotypes for genome-wide association studies (GWAS) can identify and interpret pleiotropic loci which are essential to understand pleiotropy in diseases and complex traits. Meanwhile, constructing a network based on associations between phenotypes and genotypes provides a new insight to analyze multiple phenotypes, which can explore whether phenotypes and genotypes might be related to each other at a higher level of cellular and organismal organization. In this paper, we first develop a bipartite signed network by linking phenotypes and genotypes into a Genotype and Phenotype Network (GPN). The GPN can be constructed by a mixture of quantitative and qualitative phenotypes and is applicable to binary phenotypes with extremely unbalanced case-control ratios in large-scale biobank datasets. We then apply a powerful community detection method to partition phenotypes into disjoint network modules based on GPN. Finally, we jointly test the association between multiple phenotypes in a network module and a single nucleotide polymorphism (SNP). Simulations and analyses of 72 complex traits in the UK Biobank show that multiple phenotype association tests based on network modules detected by GPN are much more powerful than those without considering network modules. The newly proposed GPN provides a new insight to investigate the genetic architecture among different types of phenotypes. Multiple phenotypes association studies based on GPN are improved by incorporating the genetic information into the phenotype clustering. Notably, it might broaden the understanding of genetic architecture that exists between diagnoses, genes, and pleiotropy.

Constructing genotype and phenotype network helps reveal disease heritability and phenome-wide association studies

“…Note that N k = N l ( k ≠ l ) if the GWAS summary statistics of the k th phenotype and l th phenotype are calculated from the same study cohort, otherwise, N k ≠ N l . For simplicity, we assume the generalized linear regression 7 , , where y ik is the k th phenotype value and X ik is the vector of covariates, for example, used to account for population stratification in the study, for the i th (1 ≤ i ≤ N k ) individual and the k th phenotype. Assuming that there are M k genetic variants in the GWAS summary statistics for the k th phenotype and g im is the genotype of the m th (1 ≤ m ≤ M k ) genetic variant taking values from 0, 1, and 2 that counts the number of copies of the minor allele.…”

“…We first define a signed bipartite GPN, 𝒢 GPN = ( Y,G, E ), where Y = { Y 1 ,…, Y K } and G ={ G 1 ,…, G M } denote two disjoint and independent sets of phenotypes and genetic variants, and E denotes the set of edges in GPN. Similar to our previous work 7 , we denote T = ( T mk ) as an M × K adjacency matrix of the denser representation of GPN, where is the weight of the edge between the m th genetic variant and the k th phenotype. F Chi (•) denotes the cumulative distribution function (CDF) of if ; if ; otherwise, sign .…”

“…Rather than simply identifying the association between a genetic variant and a specific disease, the construction of a bipartite network can reveal the integrated molecular underpinnings of diseases 5 . Therefore, a bipartite network can be used to explore whether human diseases or complex traits and the corresponding genetic variants are related to each other at a higher level of cellular and organization 6; 7 . In addition, due to many complex diseases being affected by a shared set of pleiotropic variants, the construction of a bipartite network can also be used to determine the pathobiological relationship of one disease to other diseases 5 and elucidate the complexities of genetic correlations across diseases 6 .…”

“…Over the past decade, genome-wide association studies (GWAS) have generated an impressive list of genetic variant and phenotype association pairs 8; 9 , which offer a great opportunity to establish a bipartite network connecting genetic variants and phenotypes, referred to as a genotype and phenotype network (GPN) 7 . GPN provides integrative analyses that allow for the characterization of complex relationships between genetic variants and phenotypes, which are reproducible and accurately represent biological relationships.…”

“…Phenotypes in the whole phenome can be grouped by digitized codes (e.g., ICD-10 code) to represent the common clinical factors underlying the diseases. However, the taxonomy of digitized codes is based on their etiology rather than their genetic architecture, but applying the community detection method for GPN allows us to identify network modules that provide an integrative approach to understanding the complex genetic relationships across phenotypes 7 . A network module is loosely defined as a subnetwork with high local link density so that the phenotypes within a network module share more genetic architecture across all genetic variants than phenotypes outside the network module 19; 20 .…”