Evidence that FGFRL1 contributes to congenital diaphragmatic hernia development in humans

Identifying disease-related genes is of importance for understanding molecule mechanisms diseases, as well diagnosis and treatment diseases. Many computational methods have been proposed to predict genes, but how make full use multi-source biological data enhance the ability disease-gene prediction still challenging. In this paper, we a novel method predicting by using fast network embedding (PrGeFNE), which can integrate multiple types associations related diseases genes. Specifically, first constructed heterogeneous phenotype-disease, disease-gene, protein-protein gene-GO associations; low-dimensional representation nodes extracted from algorithm. Then, dual-layer was reconstructed representation, propagation applied Through cross-validation newly added-association validation, displayed important roles different association in enhancing prediction, confirmed excellent performance PrGeFNE comparing state-of-the-art algorithms. Furthermore, developed web tool that facilitate researchers search candidate predicted PrGeFNE, along with enrichment analysis GO pathway on gene set. This may be useful investigation diseases’ molecular their experimental validations. The available at http://bioinformatics.csu.edu.cn/prgefne/.