Virtual screening aims to identify and rank compounds with drug/lead-like properties based on their affinity for the protein target. We developed a methodology that integrates structure- and ligand-based screening approaches to enhance hit rates against the TDP1 protein within a database of anthraquinone and chalcone derivatives, followed by evaluation of prioritized compounds through molecular simulations. This technique is particularly useful for training set imbalances. Four screening methods were used: QSAR, pharmacophore, shape similarity, and docking. Each method was individually trained to score compounds, and the scores were fused to create parallel Z-score fusion. The QSAR models exhibited satisfactory R2 values (0.84 to 0.75), whereas the pharmacophoric and shape similarity models demonstrated excellent performance (ROC:0.82-0.88). Docking enrichment analysis identified 6N0D as the optimal TDP1 crystal structure (ROC = 0.73). Remarkably, the consensus scoring method surpassed other screening methods, achieving the highest ROC value of 0.98. Docking screening prioritized compounds with binding modes resembling the co-crystallized ligands, whereas MMGBSA, consensus, and docking produced dynamic simulations that were as stable as the co-crystallized ligands. Additionally, the QSAR-selected compounds exhibited binding modes similar to those of commercially available TDP1 inhibitors. In this study, a strong correlation was found between the inhibitory concentrations and binding energy values of commercialized TDP1 inhibitors, indicating that the top-ranked compounds are expected to have potent inhibitory effects in the nano-/micromolar range. The results of this study establish that consensus scoring can be used as an adaptable mainstay virtual screening methodology, pending subsequent experimental validation for affirmation.Communicated by Ramaswamy H. Sarma.