In the current biomedical repositories, gene and disease identification and prediction are the essential factors for content clustering and classification models. Since, most of the biomedical databases have heterogeneous features with different levels of gene patterns. Gene identification and clustering of high dimensional patterns in cross biomedical repositories are complex and difficult to process due to noise, uncertain and missing values. In the traditional biomedical repositories, data classification algorithms are used to classify the documents using the MeSH terms or user specific keywords. These models are difficult to find the relational genes and its disease patterns in different biomedical repositories. In the proposed work, a hybrid cross gene-chemical-disease based document clustering and classification model is implemented using the deep learning framework. Experimental results proved that the proposed deep learning-based G-C-D document classification has better optimization than the existing models.