Molecular Forecaster is pleased to announce that one of our compounds has advanced to Round 2 in CACHE Challenge #6, marking our second consecutive challenge with a compound moving forward. This achievement reflects our team’s continued refinement of integrated computational approaches in drug discovery.
The Challenge
CACHE Challenge #6 focused on identifying novel inhibitors for SETDB1, an epigenetic target with a particularly challenging binding site. The Triple Tudor Domain features multiple sub-pockets, including two aromatic cages and an acetyl-binding groove, that can be targeted for effective inhibition.
Our Approach
The MFI team employed a comprehensive structure-based strategy that combined:
- Comparative analysis of 21 SETDB1 crystal structures
- Benchmarking through self-docking and cross-docking protocols
- Construction of five structure-based pharmacophore models
- Large-scale docking screens executed on local HPC and Azure cloud resources
- Expert visual scoring by our team
- High-throughput molecular dynamics rescoring and affinity ranking in partnership with HTuO
- Final candidate selection informed by integrated analysis across all methods
Figure 1: MFI’s integrated computational workflow for CACHE Challenge #6
Results
Out of the 100 compounds submitted, 76 were successfully synthesized by Enamine and tested by Conscience. One compound demonstrated micromolar binding activity and has been advanced to Round 2. Notably, this hit showed an unremarkable initial docking score but received consistent visual scoring across our team and strong affinity predictions during physics-based rescoring—a reminder that layering multiple validation methods provides greater predictive value than any single computational approach.
Next Steps
Round 2 will focus on submitting 50 close analogues by mid-January 2026 to establish structure-activity relationships and confirm the validity of the initial hit. MFI will continue our collaboration with HTuO Biosciences throughout this next phase.
This achievement highlights the effectiveness of combining structural insight, scalable computation, and physics-based refinement for rapid hit identification in challenging drug discovery programs.