Computational Chemist
Benchmark, validate, and complement in silico workflows.
Recommended reading order
4 pages- 1Getting StartedSubmitting your first retrosynthesisRun a target molecule end-to-end, from SMILES to finished routes.
- 2FeaturesFeasibility scoreThe composite 0-10 score and its five dimensions.
- 3FeaturesScoring profilesBuilt-in and custom profiles for ranking routes by weighted criteria.
- 4FeaturesConvergence explorerExplore precursors, products, and co-occurring molecules for intermediates.
Why this role uses SynovAI
Leverage SynovAI to complement computational workflows — from validating predicted targets to exploring novel disconnections.
Benchmarking retrosynthetic predictions
Submit a target molecule with a known published synthesis and compare the platform's proposed routes against the literature. Use the confidence scores and reference entries to assess how well the model recovers established chemistry — valuable for calibrating trust in predictions for novel targets.
Exploring disconnection strategies for complex architectures
For structurally complex targets (natural products, macrocycles, densely functionalized scaffolds), run multiple analyses varying step count and precedent level. "Well-Precedented" routes reveal classical strategies, while "Novel" mode surfaces unconventional disconnections that may inspire new retrosynthetic hypotheses worth investigating computationally.
Validating synthetic feasibility of computationally designed molecules
After a virtual screening or generative chemistry campaign produces candidate structures, quickly assess synthetic accessibility by submitting each hit. Routes with high confidence and commercially available starting materials indicate tractable targets — helping prioritize which candidates to advance to synthesis.