| Complex diseases are often caused by the combined effects of multiple genetic factors,with a complex genetic mechanism that is difficult to accurately identify and explain.Genome-wide association study aim to reveal the genetic mechanism behind complex diseases by identifying the association between single nucleotide polymorphism(SNP)local and complex diseases.However,with the increase of data scale,the interaction patterns between SNPs also become increasingly complex.Therefore,effectively identifying SNP interaction patterns has become one of the research hotspots in the field of bioinformatics,and is of great significance to study the pathogenesis of complex diseases and formulate later treatment plans.Based on the simulated disease models and real complex disease data,this dissertation aims to identify SNP interaction patterns related to the occurrence and development of complex diseases using an optimized artificial bee colony(ABC)search method.The specific research contents are as follows:(1)Aiming at the two-order SNP interaction patterns recognition problem,a multi-objective ABC method based on scale-free network is proposed.Firstly,the complex network structure is used to guide the bee colony to update iteratively.The low degree-degree coefficient characteristic of scale-free network is used to make it easier for the low-quality solution approach to the highquality solution.Secondly,the multi-objective strategy and opposition-based learning strategy are used in the search process,which can improve the recognition ability of the method and effectively ensure the diversity of the population.Finally,the method is experimented in the simulated datasets and the real age-related macular degeneration disease dataset,and shows good recognition performance.(2)Aiming at the recognition problem of high-order SNP interaction patterns with marginal effects,an adaptive multi-population ABC method based on contribution is proposed.Firstly,the contribution-based feature space grouping strategy is used to improve the ability of the method to deal with large-scale data.Secondly,the search reinforcement-based update strategy is used to automatically adjust the positive and negative selection of food sources,thereby enhancing the search ability of ABC.Then,the reasonable utilization of computing resources is realized through the adaptive iteration strategy.Finally,experiments on the simulated datasets and the real disease datasets demonstrate that the proposed method can effectively identify the high-order SNP interactions with marginal effects.(3)Aiming at the recognition problem of high-order SNP interaction patterns without marginal effects,an adaptive ABC method based on self-adjusting random grouping is proposed.Firstly,the random grouping strategy based on self-adjustment is used to improve the probability of diseasecasing SNP combinations being grouped into the same subset.Secondly,through the adaptive iteration strategy based on variance,the reasonable utilization of computing resources can be realized while ensuring population diversity.Then,the solution retention strategy is used to optimize the dependence on control parameters during the scout bee stage.Finally,the proposed method is applied to the simulation datasets and the real complex disease dataset,which proves that it can effectively identify the high-order SNP interactions without marginal effects.The experimental results show that the methods proposed in this dissertation have more advantages in identifying interaction patterns than similar recognition methods,and can more stably and effectively identify genetic interaction patterns related to complex diseases. |
