MIPD: Molecules, Imagings, and Clinical Phenotype Integrated Database

Abstract

Due to tumor heterogeneity, a subset of patients fails to benefit from current treatment strategies. However, an integrated analysis of imaging features, genetic molecules, and clinical phenotypes can characterize tumor heterogeneity, enabling the development of more personalized treatment approaches. Despite its potential, cross-modal databases remain underexplored. To address this gap, we established a comprehensive database encompassing 9965 genes, 5449 proteins, 1121 metabolites, 283 pathways, 854 imaging features, and 73 clinical factors from colorectal cancer patients. This database identifies significantly distinct molecules and imaging features associated with clinical phenotypes and provides survival analysis based on these features. Additionally, it offers genetic molecule annotations, comparative expression levels between tumor and normal tissues, imaging features linked to genetic molecules, and imaging-based models for predicting gene expression levels. Furthermore, the database highlights correlations between genetic molecules, clinical factors, and imaging features. In summary, we present MIPD (Molecules, Imaging, and Clinical Phenotype Correlation Database), a user-friendly, interactive, and specialized platform accessible at http://corgenerf.com. MIPD facilitates the interpretability of cross-modal data by providing query, browse, search, visualization, and download functionalities, thereby offering a valuable resource for advancing precision medicine in colorectal cancer. Database URL: http://corgenerf.

Figure 1.

Database content and construction of MIPD. (a) Data collection process in the MIPD. (b) Construction of MIPD. (c) User interface of the MIPD, supporting browsing, searching, and downloading of analysis results.

Figure 2.

The main functions and usages of MIPD. Input, output, and contents information of MIPD.

Figure 3.

Overview of MIPD knowledgebase. (a) Homepage and keyword search interface. (b) Example of searching for the clinical factor ‘OS’. (c) Example of searching for the molecule ‘CD4’. (d) Example of searching for the imaging feature ‘original_shape_Flatness’.