MOSES: A New Approach to Integrate Interactome Topology and Functional Features for Disease Gene Prediction

Disease gene prediction is to date one of the main computational challenges of precision medicine. It is still uncertain if disease genes have unique functional properties that distinguish them from other non-disease genes or, from a network perspective, if they are located randomly in the interactome or show specific patterns in the network topology. In this study, we propose a new method for disease gene prediction based on the use of biological knowledge-bases (gene-disease associations, genes functional annotations, etc.) and interactome network topology. The proposed algorithm called MOSES is based on the definition of two somewhat opposing sets of genes both disease-specific from different perspectives: warm seeds (i.e., disease genes obtained from databases) and cold seeds (genes far from the disease genes on the interactome and not involved in their biological functions). The application of MOSES to a set of 40 diseases showed that the suggested putative disease genes are significantly enriched in their reference disease. Reassuringly, known and predicted disease genes together, tend to form a connected network module on the human interactome, mitigating the scattered distribution of disease genes which is probably due to both the paucity of disease-gene associations and the incompleteness of the interactome.

HyMM: Hybrid method for disease-gene prediction by integrating multiscale module structures

Motivation: Identifying disease-related genes is important for the study of human complex diseases. Module structures or community structures are ubiquitous in biological networks. Although the modular nature of human diseases can provide useful insights, the mining of information hidden in multiscale module structures has received less attention in disease-gene prediction. Results: We propose a hybrid method, HyMM, to predict disease-related genes more effectively by integrating the information from multiscale module structures. HyMM consists of three key steps: extraction of multiscale modules, gene rankings based on multiscale modules and integration of multiple gene rankings. The statistical analysis of multiscale modules extracted by three multiscale-module-decomposition algorithms (MO, AS and HC) shows that the functional consistency of the modules gradually improves as the resolution increases. This suggests the existence of different levels of functional relationships in the multiscale modules, which may help reveal disease-gene associations. We display the effectiveness of multiscale module information in the disease-gene prediction and confirm the excellent performance of HyMM by 5-fold cross-validation and independent test. Specifically, HyMM with MO can more effectively enhance the ability of disease-gene prediction; HyMM (MO, RWR) and HyMM (MO, RWRH) are especially preferred due to their excellent comprehensive performance, and HyMM (AS, RWRH) is also good choice due to its local performance. We anticipate that this work could provide useful insights for disease-module analysis and disease-gene prediction based on multi-scale module structures. Availability: https://github.com/xiangju0208/HyMM

Recent advances in network-based methods for disease gene prediction

Disease–gene association through genome-wide association study (GWAS) is an arduous task for researchers. Investigating single nucleotide polymorphisms that correlate with specific diseases needs statistical analysis of associations. Considering the huge number of possible mutations, in addition to its high cost, another important drawback of GWAS analysis is the large number of false positives. Thus, researchers search for more evidence to cross-check their results through different sources. To provide the researchers with alternative and complementary low-cost disease–gene association evidence, computational approaches come into play. Since molecular networks are able to capture complex interplay among molecules in diseases, they become one of the most extensively used data for disease–gene association prediction. In this survey, we aim to provide a comprehensive and up-to-date review of network-based methods for disease gene prediction. We also conduct an empirical analysis on 14 state-of-the-art methods. To summarize, we first elucidate the task definition for disease gene prediction. Secondly, we categorize existing network-based efforts into network diffusion methods, traditional machine learning methods with handcrafted graph features and graph representation learning methods. Thirdly, an empirical analysis is conducted to evaluate the performance of the selected methods across seven diseases. We also provide distinguishing findings about the discussed methods based on our empirical analysis. Finally, we highlight potential research directions for future studies on disease gene prediction.

Machine learning-based approaches for disease gene prediction

Disease gene prediction is an essential issue in biomedical research. In the early days, annotation-based approaches were proposed for this problem. With the development of high-throughput technologies, interaction data between genes/proteins have grown quickly and covered almost genome and proteome; thus, network-based methods for the problem become prominent. In parallel, machine learning techniques, which formulate the problem as a classification, have also been proposed. Here, we firstly show a roadmap of the machine learning-based methods for the disease gene prediction. In the beginning, the problem was usually approached using a binary classification, where positive and negative training sample sets are comprised of disease genes and non-disease genes, respectively. The disease genes are ones known to be associated with diseases; meanwhile, non-disease genes were randomly selected from those not yet known to be associated with diseases. However, the later may contain unknown disease genes. To overcome this uncertainty of defining the non-disease genes, more realistic approaches have been proposed for the problem, such as unary and semi-supervised classification. Recently, more advanced methods, including ensemble learning, matrix factorization and deep learning, have been proposed for the problem. Secondly, 12 representative machine learning-based methods for the disease gene prediction were examined and compared in terms of prediction performance and running time. Finally, their advantages, disadvantages, interpretability and trust were also analyzed and discussed.