Extracellular matrices are essential for cellular and organismal function. Recent genome-wide and phenome-wide association studies started to reveal a broad spectrum of phenotypes associated with genetic variants. However, the phenome or spectrum of all phenotypes associated with genetic variants in extracellular matrix genes is unknown. Here, we analyzed over two million recorded genotype-to-phenotype relationships across multiple species to define their extracellular matrix phenomes. By using the previously defined matrisomes of humans, mice, zebrafish, Drosophila, and C. elegans, we found that the extracellular matrix phenome comprises of 3-10% of the entire phenome. Collagens (COL1A1, COL2A1) and fibrillin (FBN1) are each associated with >150 distinct phenotypes in humans, whereas collagen COL4A1, Wnt- and sonic hedgehog (shh) signaling are predominantly associated in other species. We determined the phenotypic fingerprints of matrisome genes and found that MSTN, CTSD, LAMB2, HSPG2, and COL11A2 and their corresponding orthologues have the most phenotypes across species. Out of the 42,551 unique matrisome genotype-to-phenotype relationships across the five species with defined matrisomes, we have constructed interaction networks to identify the underlying molecular components connecting with orthologues phenotypes and with novel phenotypes. Thus, our networks provide a framework to predict unassessed phenotypes and their potential underlying molecular interactions. These frameworks inform on matrisome genotype-to-phenotype relationships and potentially provide a sophisticated choice of biological model system to study human phenotypes and diseases.
Keywords: Collagen; Data mining; Extracellular matrix; Genotype-to-phenotype; Matrisome; Phenome.
© 2020 The Author(s).