Untitled

Funder

Organizational Unit

Publications

Biomedical knowledge graph embeddings for personalized medicine: Predicting disease‐gene associations

Publication . Vilela, Joana; Asif, Muhammad; Marques, Ana Rita; Santos, João Xavier; Rasga, Célia; Vicente, Astrid; Martiniano, Hugo

Personalized medicine is a concept that has been subject of increasing interest in medical research and practice in the last few years. However, significant challenges stand in the way of practical implementations, namely in regard to extracting clinically valuable insights from the vast amount of biomedical knowledge generated in the last few years. Here, we describe an approach that uses Knowledge Graph Embedding (KGE) methods on a biomedical Knowledge Graph (KG) as a path to reasoning over the wealth of information stored in publicly accessible databases. We built a Knowledge Graph using data from DisGeNET and GO, containing relationships between genes, diseases and other biological entities. The KG contains 93,657 nodes of 5 types and 1,705,585 relationships of 59 types. We applied KGE methods to this KG, obtaining an excellent performance in predicting gene-disease associations (MR 0.13, MRR 0.96, HITS@1 0.93, HITS@3 0.99, and HITS@10 0.99). The optimal hyperparameter set was used to predict all possible novel gene-disease associations. An in-depth analysis of novel gene-disease predictions for disease terms related to Autism Spectrum Disorder (ASD) shows that this approach produces predictions consistent with known candidate genes and biological pathways and yields relevant insights into the biology of this paradigmatic complex disorder.

2022-11-20Journal article

Open access

Evidence for an association of prenatal exposure to particulate matter with clinical severity of Autism Spectrum Disorder

Publication . Santos, João Xavier; Sampaio, Pedro; Rasga, Célia; Martiniano, Hugo; Faria, Clarissa; Café, Cátia; Oliveira, Alexandra; Duque, Frederico; Oliveira, Guiomar; Sousa, Lisete; Nunes, Ana; Moura Vicente, Astrid

Early-life exposure to air pollutants, including ozone (O3), particulate matter (PM2.5 or PM10, depending on diameter of particles), nitrogen dioxide (NO2) and sulfur dioxide (SO2) has been suggested to contribute to the etiology of Autism Spectrum Disorder (ASD). In this study, we used air quality monitoring data to examine whether mothers of children with ASD were exposed to high levels of air pollutants during critical periods of pregnancy, and if higher exposure levels may lead to a higher clinical severity in their offspring. We used public data from the Portuguese Environment Agency to estimate exposure to these pollutants during the first, second and third trimesters of pregnancy, full pregnancy and first year of life of the child, for 217 subjects with ASD born between 2003 and 2016. These subjects were stratified in two subgroups according to clinical severity, as defined by the Autism Diagnostic Observational Schedule (ADOS). For all time periods, the average levels of PM2.5, PM10 and NO2 to which the subjects were exposed were within the admissible levels defined by the European Union. However, a fraction of these subjects showed exposure to levels of PM2.5 and PM10 above the admissible threshold. A higher clinical severity was associated with higher exposure to PM2.5 (p = 0.001), NO2 (p = 0.011) and PM10 (p = 0.041) during the first trimester of pregnancy, when compared with milder clinical severity. After logistic regression, associations with higher clinical severity were identified for PM2.5 exposure during the first trimester (p = 0.002; OR = 1.14, 95%CI: 1.05–1.23) and full pregnancy (p = 0.04; OR = 1.07, 95%CI: 1.00–1.15) and for PM10 (p = 0.02; OR = 1.07, 95%CI: 1.01–1.14) exposure during the third trimester. Exposure to PM is known to elicit neuropathological mechanisms associated with ASD, including neuroinflammation, mitochondrial disruptions, oxidative stress and epigenetic changes. These results offer new insights on the impact of earlylife exposure to PM in ASD clinical severity.

2023-04-05Journal article

Open access

A Role for Gene-Environment Interactions in Autism Spectrum Disorder Is Supported by Variants in Genes Regulating the Effects of Exposure to Xenobiotics

Publication . Santos, João Xavier; Rasga, Célia; Marques, Ana Rita; Martiniano, Hugo; Asif, Muhammad; Vilela, Joana; Oliveira, Guiomar; Sousa, Lisete; Nunes, Ana; Vicente, Astrid M.

Heritability estimates support the contribution of genetics and the environment to the etiology of Autism Spectrum Disorder (ASD), but a role for gene-environment interactions is insufficiently explored. Genes involved in detoxification pathways and physiological permeability barriers (e.g., blood-brain barrier, placenta and respiratory airways), which regulate the effects of exposure to xenobiotics during early stages of neurodevelopment when the immature brain is extremely vulnerable, may be particularly relevant in this context. Our objective was to identify genes involved in the regulation of xenobiotic detoxification or the function of physiological barriers (the XenoReg genes) presenting predicted damaging variants in subjects with ASD, and to understand their interaction patterns with ubiquitous xenobiotics previously implicated in this disorder. We defined a panel of 519 XenoReg genes through literature review and database queries. Large ASD datasets were inspected for in silico predicted damaging Single Nucleotide Variants (SNVs) (N = 2,674 subjects) or Copy Number Variants (CNVs) (N = 3,570 subjects) in XenoReg genes. We queried the Comparative Toxicogenomics Database (CTD) to identify interaction pairs between XenoReg genes and xenobiotics. The interrogation of ASD datasets for variants in the XenoReg gene panel identified 77 genes with high evidence for a role in ASD, according to pre-specified prioritization criteria. These include 47 genes encoding detoxification enzymes and 30 genes encoding proteins involved in physiological barrier function, among which 15 are previous reported candidates for ASD. The CTD query revealed 397 gene-environment interaction pairs between these XenoReg genes and 80% (48/60) of the analyzed xenobiotics. The top interacting genes and xenobiotics were, respectively, CYP1A2, ABCB1, ABCG2, GSTM1, and CYP2D6 and benzo-(a)-pyrene, valproic acid, bisphenol A, particulate matter, methylmercury, and perfluorinated compounds. Individuals carrying predicted damaging variants in high evidence XenoReg genes are likely to have less efficient detoxification systems or impaired physiological barriers. They can therefore be particularly susceptible to early life exposure to ubiquitous xenobiotics, which elicit neuropathological mechanisms in the immature brain, such as epigenetic changes, oxidative stress, neuroinflammation, hypoxic damage, and endocrine disruption. As exposure to environmental factors may be mitigated for individuals with risk variants, this work provides new perspectives to personalized prevention and health management policies for ASD.

2022-05-19Journal article

Open access