STAT3In

STAT3In predicts small-molecule inhibitors of the Signal Transducer and Activator of Transcription 3 (STAT3) signaling pathway to support identification and repurposing of therapeutics targeting IL6-mediated inflammatory responses.

Key Features:

• Dataset Utilization: Models were developed from a dataset of 1564 known small-molecule STAT3 inhibitors and 1671 non-inhibitors used for training, testing, and evaluation.
• Chemical Descriptor Analysis: Computed two-dimensional (2-D) and three-dimensional (3-D) chemical descriptors for each compound.
• Model Development: Constructed descriptor-based, fingerprint (FP)-based, and hybrid-feature models, with 2-D and 3-D descriptor models achieving AUCs of 0.84 and 0.73 respectively, FP-based models achieving AUC 0.86 and validation accuracy 78.70%, and hybrid models reaching AUC 0.87 on the validation dataset.
• Identification of Potential Therapeutics: Predictions identified FDA-approved drugs including Tamoxifen and Perindopril as candidate STAT3 inhibitors.

Scientific Applications:

• Drug discovery and repurposing: Prioritize and nominate small molecules and approved drugs from chemical libraries as candidate STAT3 inhibitors for experimental validation.
• Inflammation and COVID-19 research: Support identification of inhibitors targeting IL6-mediated STAT3 signaling implicated in proinflammatory responses and cytokine storms in severe COVID-19 cases.

Methodology:

Compiled a dataset of 1564 STAT3 inhibitors and 1671 non-inhibitors, computed 2-D and 3-D chemical descriptors and molecular fingerprints, and developed and evaluated descriptor-based, FP-based, and hybrid predictive models.