多数野生植物可经过简单的处理作为草食畜禽的饲料。因此内蒙古野生植物的开发与利用将有利于内蒙古畜牧业的发展。蛋白质是生命活动的主要承担者，参与生命的各种活动。对这些植物蛋白质功能的研究将加快生物学家全面而深刻的了解蛋白质的分子机制，了解这些植物的生物特性及营养价值。鉴于目前对内蒙古野生植物蛋白质功能研究较少，预测模型匮乏，故本项目对内蒙古野生植物蛋白质功能预测研究。本项目首先研究蛋白质特征提取方法。为了提高预测精度，增加特征信息，故基于多源数据提取特征。其次研究特征选择方法。基于蛋白质相互作用网络改进ReliefF，使得此方法可以解决多标签分类中的特征选择问题。最后研究分类模型。提出基于深度卷积神经网络的蛋白质功能分类模型，设置卷积层、池化层、输出层等，优化参数，提高分类精度。利用分类器对内蒙古野生植物蛋白质功能进行预测，分析蛋白质功能与内蒙古野生植物的生物特性之间的关系。

Most wild plants can be simply processed as feed for herbivorous livestock. Therefore, the development and utilization of wild plants in Inner Mongolia will be beneficial to the development of animal husbandry in Inner Mongolia. Protein is the main bearer of life activities and participates in various activities of life. The study of the protein function for these plants will help biologists to understand profoundly the molecular mechanisms of proteins, the biological characteristics and nutritional value of these plants. In view of the lack of the research on protein function of wild plants in Inner Mongolia and predictive models, this project researches the prediction of protein function for wild plant in Inner Mongolia. This project first studies the method of protein feature extraction. In order to improve the prediction accuracy and increase the feature information, features are extracted based on the multi-source data. Protein interaction networks will be constructed, and functional features will be extracted from the protein interaction network using a random forest with restart. Next, this project studies the feature selection method. ReliefF is improved based on the protein interaction network, such that the new method can solve the feature selection of the multi-label classification problem. Finally, the project studies the classification model. It will propose a model for classification of protein function based on deep convolution neural network, set the convolutional layer, pooling layer and output layer, optimize parameters and improve classification accuracy. The classifier was used to predict the function of protein for wild plant in Inner Mongolia. It analyzes the relationship between its protein function and the biological characteristics of wild plants in Inner Mongolia.